diff --git a/ruins/apps/weather.py b/ruins/apps/weather.py
index 2c0b9b9..56b395e 100644
--- a/ruins/apps/weather.py
+++ b/ruins/apps/weather.py
@@ -1,3 +1,4 @@
+from typing import List, Callable, Tuple
 import streamlit as st
 import xarray as xr     # TODO: these references should be moved to DataManager
 import pandas as pd     # TODO: these references should be moved to DataManager
@@ -6,26 +7,13 @@
 
 from ruins.plotting import plt_map, kde, yrplot_hm
 from ruins import components
+from ruins.core import build_config, DataManager, Config
+from ruins.core.cache import partial_memoize
 
 
 
 ####
 # OLD STUFF
-# 
-# TODO: replace with DataManager
-def load_alldata():
-    weather = xr.load_dataset('data/weather.nc')
-    climate = xr.load_dataset('data/cordex_coast.nc')
-
-    # WARNING - bug fix for now:
-    # 'HadGEM2-ES' model runs are problematic and will be removed for now
-    # The issue is with the timestamp and requires revision of the ESGF reading routines
-    kys = [s for s in list(climate.keys()) if 'HadGEM2-ES' not in s] #remove all entries of HadGEM2-ES (6 entries)
-    climate = climate[kys]
-
-    return weather, climate
-
-
 def applySDM(wdata, data, meth='rel', cdf_threshold=0.9999999, lower_limit=0.1):
     '''apply structured distribution mapping to climate data and return unbiased version of dataset'''
     from sdm import SDM
@@ -75,7 +63,15 @@ def climate_indi(ts, indi='Summer days (Tmax ≥ 25°C)'):
 
 # TODO: document + signature
 # TODO: extract plotting
-def climate_indices(weather: xr.Dataset, climate: xr.Dataset, stati='coast', cliproj=True):
+def climate_indices(dataManager: DataManager, config: Config):
+    # get data
+    weather = dataManager['weather'].read()
+    climate = dataManager['cordex_coast'].read()
+
+    # get the relevant settings
+    stati = config.get('selected_station', 'coast')
+
+
     cindi = ['Ice days (Tmax < 0°C)', 'Frost days (Tmin < 0°C)', 'Summer days (Tmax ≥ 25°C)', 'Hot days (Tmax ≥ 30°C)','Tropic nights (Tmin ≥ 20°C)', 'Rainy days (Precip ≥ 1mm)']
     ci_topic = st.selectbox('Select Index:', cindi)
 
@@ -92,12 +88,12 @@ def climate_indices(weather: xr.Dataset, climate: xr.Dataset, stati='coast', cli
     w1 = weather[stati].sel(vars=vari).to_dataframe().dropna()
     w1.columns = ['bla', vari]
 
-    plt.figure(figsize=(10,2.5))
+    fig = plt.figure(figsize=(10,2.5))
     wi = climate_indi(w1, ci_topic).astype(int)
     wi.plot(style='.', color='steelblue', label='Coast weather')
     wi.rolling(10, center=True).mean().plot(color='steelblue', label='Rolling mean\n(10 years)')
 
-    if cliproj:
+    if config['include_climate']:
         c1 = climate.sel(vars=vari).to_dataframe()
         c1 = c1[c1.columns[c1.columns != 'vars']]
         c2 = applySDM(w1[vari], c1, meth=meth)
@@ -127,7 +123,7 @@ def climate_indices(weather: xr.Dataset, climate: xr.Dataset, stati='coast', cli
     plt.legend(ncol=2)
     plt.ylabel('Number of days')
     plt.title(ci_topic)
-    st.pyplot()
+    st.pyplot(fig)
 
     if ci_topic == 'Ice days (Tmax < 0°C)':
         st.markdown('''Number of days in one year which persistently remain below 0°C air temperature.''')
@@ -144,178 +140,246 @@ def climate_indices(weather: xr.Dataset, climate: xr.Dataset, stati='coast', cli
     return
 
 
-def weather_explorer(w_topic: str):
-    weather, climate = load_alldata()
-    #weather = load_data('Weather')
+def data_select(dataManager: DataManager, config: Config, container=st) -> None:
+    """Create the user interface to control the data view
+    """
+    # get a station list
+    weather = dataManager['weather'].read()
+    station_list = list(weather.keys()) # TODO station names krummhoern, coast, inland, niedersachsen?
+    selected_station = container.selectbox('Select station/group (see map in sidebar for location):', station_list)
 
-    #aspects = ['Annual', 'Monthly', 'Season']
-    #w_aspect = st.sidebar.selectbox('Temporal aggegate:', aspects)
+    # select a temporal aggregation
+    aggregations = config.get('temporal_aggregations', ['Annual', 'Monthly'])
+    temp_agg = container.selectbox('Select temporal aggregation:', aggregations)
+
+    # include climate projections
+    include_climate = container.checkbox('Include climate projections (for coastal region)', value=False)
+
+    # add settings
+    st.session_state.selected_station = selected_station
+    st.session_state.temp_agg = temp_agg
+    st.session_state.include_climate = include_climate
+
+
+@partial_memoize(hash_names=['name', 'station', 'variable', 'time', '_filter'])
+def _reduce_weather_data(dataManager: DataManager, name: str, variable: str, time: str, station: str = None, _filter: dict = None) -> pd.DataFrame:
+    # get weather data
+    arr: xr.Dataset = dataManager[name].read()
+
+    if _filter is not None:
+        arr = arr.filter_by_attrs(**_filter)
+
+    if station is None:
+        base = arr
+    else:
+        base = arr[station]
 
-    #cliproj = st.sidebar.checkbox('add climate projections',False)
+    # reduce to station and variable
+    reduced = base.sel(vars=variable).resample(time=time)
 
+    if variable == 'Tmax':
+        df = reduced.max(dim='time').to_dataframe()
+    elif variable == 'Tmin':
+        df = reduced.min(dim='time').to_dataframe()
+    else:
+        df = reduced.mean(dim='time').to_dataframe()
+    
+    if station is None:
+        return df.loc[:, df.columns != 'vars']
+    else:
+        return df[station]       
+
+
+def warming_data_plotter(dataManager: DataManager, config: Config):
+    weather: xr.Dataset = dataManager['weather'].read()
+    climate = dataManager['cordex_coast'].read()
     statios = list(weather.keys())
-    stat1 = st.selectbox('Select station/group (see map in sidebar for location):', statios)
+    stat1 = config['selected_station']
 
-    aspects = ['Annual', 'Monthly']  # , 'Season']
-    w_aspect = st.selectbox('Select temporal aggegate:', aspects)
+    # TODO refactor in data-aggregator and data-plotter for different time frames
 
-    cliproj = st.checkbox('add climate projections (for coastal region)',False)
-    if cliproj:
-        plt_map(stat1, 'CORDEX')
-        st.sidebar.markdown(
-            '''Map with available stations (<span style="color:blue">blue dots</span>) and selected reference station (<span style="color:magenta">magenta highlight</span>). The climate model grid is given in <span style="color:orange">orange</span> with the selected references as filled dots.''',
-            unsafe_allow_html=True)
+    # ----
+    # data-aggregator controls
+    navi_vars = ['Maximum Air Temperature', 'Mean Air Temperature', 'Minimum Air Temperature']
+    navi_var = st.sidebar.radio("Select variable:", options=navi_vars)
+    if navi_var[:4] == 'Mini':
+        vari = 'Tmin'
+        ag = 'min'
+    elif navi_var[:4] == 'Maxi':
+        vari = 'Tmax'
+        ag = 'max'
     else:
-        plt_map(stat1)
-        st.sidebar.markdown(
-            '''Map with available stations (<span style="color:blue">blue dots</span>) and selected reference station (<span style="color:magenta">magenta highlight</span>).''',
-            unsafe_allow_html=True)
+        vari = 'T'
+        ag = 'mean'
+
+    # controls end
+    # ----
+
+    # TODO: this produces a slider but also needs some data caching
+    if config['temp_agg'] == 'Annual':
+        wdata = _reduce_weather_data(dataManager, name='weather', station=config['selected_station'], variable=vari, time='1Y')
+        allw = _reduce_weather_data(dataManager, name='weather', variable=vari, time='1Y')
+
+        dataLq = float(np.floor(allw.min().quantile(0.22)))
+        datamin = float(np.min([dataLq, np.round(allw.min().min(), 1)]))
+        
+        if config['include_climate']:
+            rcps = ['rcp26', 'rcp45', 'rcp85']
+            rcp = st.selectbox(
+                'RCP (Mean over all projections will be shown. For more details go to section "Climate Projections"):',
+                rcps)
+
+            data = _reduce_weather_data(dataManager, name='cordex_coast', variable=vari, time='1Y', _filter=dict(RCP=rcp))
+            data_ub = applySDM(wdata, data, meth='abs')
+
+            dataUq = float(np.ceil(data_ub.max().quantile(0.76)))
+            datamax = float(np.max([dataUq, np.round(data_ub.max().max(), 1)]))
+        else:
+            dataUq = float(np.ceil(allw.max().quantile(0.76)))
+            datamax = float(np.max([dataUq,np.round(allw.max().max(), 1)]))
+
+        datarng = st.slider('Adjust data range on x-axis of plot:', min_value=datamin, max_value=datamax, value=(dataLq, dataUq), step=0.1, key='drangew')
 
-    if w_topic == 'Warming':
-
-        navi_vars = ['Maximum Air Temperature', 'Mean Air Temperature', 'Minimum Air Temperature']
-        navi_var = st.sidebar.radio("Select variable:", options=navi_vars)
-        if navi_var[:4] == 'Mini':
-            vari = 'Tmin'
-            afu = np.min
-            ag = 'min'
-        elif navi_var[:4] == 'Maxi':
-            vari = 'Tmax'
-            afu = np.max
-            ag = 'max'
+        # -------------------
+        # start plotting plot
+        if config['include_climate']:
+            fig, ax = kde(wdata, data_ub.mean(axis=1), split_ts=3)
         else:
-            vari = 'T'
-            afu = np.mean
-            ag = 'mean'
-
-        if w_aspect == 'Annual':
-            wdata = weather[stat1].sel(vars=vari).resample(time='1Y').apply(afu).to_dataframe()[stat1]
-            wdata = wdata[~np.isnan(wdata)]
-            allw = weather.sel(vars=vari).resample(time='1Y').apply(afu).to_dataframe().iloc[:, 1:]
-
-            dataLq = float(np.floor(allw.min().quantile(0.22)))
-            datamin = float(np.min([dataLq, np.round(allw.min().min(), 1)]))
-            if cliproj:
-                rcps = ['rcp26', 'rcp45', 'rcp85']
-                rcp = st.selectbox(
-                    'RCP (Mean over all projections will be shown. For more details go to section "Climate Projections"):',
-                    rcps)
-
-                data = climate.filter_by_attrs(RCP=rcp).sel(vars=vari).resample(time='1Y').apply(afu).to_dataframe()
-                data = data[data.columns[data.columns != 'vars']]
-                data_ub = applySDM(wdata, data, meth='abs')
+            fig, ax = kde(wdata, split_ts=3)
 
-                dataUq = float(np.ceil(data_ub.max().quantile(0.76)))
-                datamax = float(np.max([dataUq, np.round(data_ub.max().max(), 1)]))
-            else:
-                dataUq = float(np.ceil(allw.max().quantile(0.76)))
-                datamax = float(np.max([dataUq,np.round(allw.max().max(), 1)]))
+        ax.set_title(stat1 + ' Annual ' + navi_var)
+        ax.set_xlabel('T (°C)')
+        ax.set_xlim(datarng[0],datarng[1])
+        st.pyplot(fig)
 
-            datarng = st.slider('Adjust data range on x-axis of plot:', min_value=datamin, max_value=datamax, value=(dataLq, dataUq), step=0.1, key='drangew')
+        sndstat = st.checkbox('Show second station for comparison')
 
-            if cliproj:
-                ax = kde(wdata, data_ub.mean(axis=1), split_ts=3)
-            else:
-                ax = kde(wdata, split_ts=3)
+        if sndstat:
+            stat2 = st.selectbox('Select second station:', [x for x in statios if x != config['selected_station']])
+            wdata2 = _reduce_weather_data(dataManager, name='weather', station=stat2, variable=vari, time='1Y')
 
-            ax.set_title(stat1 + ' Annual ' + navi_var)
-            ax.set_xlabel('T (°C)')
-            ax.set_xlim(datarng[0],datarng[1])
-            st.pyplot()
+            fig, ax2 = kde(wdata2, split_ts=3)
+            ax2.set_title(stat2 + ' Annual ' + navi_var)
+            ax2.set_xlabel('T (°C)')
+            ax2.set_xlim(datarng[0],datarng[1])
+            st.pyplot(fig)
 
-            sndstat = st.checkbox('Show second station for comparison')
+        # Re-implement this as a application wide service
+        # expl_md = read_markdown_file('explainer/stripes.md')
+        # st.markdown(expl_md, unsafe_allow_html=True)
 
-            if sndstat:
-                stat2 = st.selectbox('Select second station:', [x for x in statios if x != stat1])
-                wdata2 = weather[stat2].sel(vars=vari).resample(time='1Y').apply(afu).to_dataframe()[stat2]
+    elif config['temp_agg'] == 'Monthly':
+        wdata = _reduce_weather_data(dataManager, name='weather', station=config['selected_station'], variable=vari, time='1M')
 
-                ax2 = kde(wdata2, split_ts=3)
-                ax2.set_title(stat2 + ' Annual ' + navi_var)
-                ax2.set_xlabel('T (°C)')
-                ax2.set_xlim(datarng[0],datarng[1])
-                st.pyplot()
+        ref_yr = st.slider('Reference period for anomaly calculation:', min_value=int(wdata.index.year.min()), max_value=2020,value=(max(1980, int(wdata.index.year.min())), 2000))
 
-            # Re-implement this as a application wide service
-            # expl_md = read_markdown_file('explainer/stripes.md')
-            # st.markdown(expl_md, unsafe_allow_html=True)
+        if config['include_climate']:
+            rcps = ['rcp26', 'rcp45', 'rcp85']
+            rcp = st.selectbox('RCP (Mean over all projections will be shown. For more details go to section "Climate Projections"):', rcps)
 
-        elif w_aspect == 'Monthly':
-            wdata = weather[stat1].sel(vars=vari).resample(time='1M').apply(afu).to_dataframe()[stat1]
-            wdata = wdata[~np.isnan(wdata)]
-            ref_yr = st.slider('Reference period for anomaly calculation:', min_value=int(wdata.index.year.min()), max_value=2020,value=(max(1980, int(wdata.index.year.min())), 2000))
+            data = _reduce_weather_data(dataManager, name='cordex_coast', variable=vari, time='1M', _filter=dict(RCP=rcp))
 
-            if cliproj:
-                rcps = ['rcp26', 'rcp45', 'rcp85']
-                rcp = st.selectbox('RCP (Mean over all projections will be shown. For more details go to section "Climate Projections"):', rcps)
+            #ub = st.sidebar.checkbox('Apply SDM bias correction',True)
+            ub = True # simplify here and automatically apply bias correction
 
-                data = climate.filter_by_attrs(RCP=rcp).sel(vars=vari).resample(time='1M').apply(afu).to_dataframe()
-                data = data[data.columns[data.columns != 'vars']]
+            if ub:
+                data_ub = applySDM(wdata, data, meth='abs')
+                fig = yrplot_hm(pd.concat([wdata.loc[wdata.index[0]:data.index[0] - pd.Timedelta('1M')], data_ub.mean(axis=1)]),ref_yr, ag, li=2006)
+            else:
+                fig = yrplot_hm(pd.concat([wdata.loc[wdata.index[0]:data.index[0] - pd.Timedelta('1M')], data.mean(axis=1)]), ref_yr, ag, li=2006)
+
+            plt.title(stat1 + ' ' + navi_var + ' anomaly to ' + str(ref_yr[0]) + '-' + str(ref_yr[1]))
+            st.pyplot(fig)
 
-                #ub = st.sidebar.checkbox('Apply SDM bias correction',True)
-                ub = True # simplify here and automatically apply bias correction
 
-                if ub:
-                    data_ub = applySDM(wdata, data, meth='abs')
-                    yrplot_hm(pd.concat([wdata.loc[wdata.index[0]:data.index[0] - pd.Timedelta('1M')], data_ub.mean(axis=1)]),ref_yr, ag, li=2006)
-                else:
-                    yrplot_hm(pd.concat([wdata.loc[wdata.index[0]:data.index[0] - pd.Timedelta('1M')], data.mean(axis=1)]), ref_yr, ag, li=2006)
+        # TODO: break up this as well
+        else:
+            fig = yrplot_hm(wdata,ref_yr,ag)
+            plt.title(stat1 + ' ' + navi_var + ' anomaly to ' + str(ref_yr[0]) + '-' + str(ref_yr[1]))
+            st.pyplot(fig)
 
-                plt.title(stat1 + ' ' + navi_var + ' anomaly to ' + str(ref_yr[0]) + '-' + str(ref_yr[1]))
-                st.pyplot()
+            sndstat = st.checkbox('Compare to a second station?')
 
+            if sndstat:
+                stat2 = st.selectbox('Select second station:', [x for x in statios if x != stat1])
+                data2 = _reduce_weather_data(dataManager, name='weather', station=stat2, variable=vari, time='1M')
+
+                ref_yr2 = list(ref_yr)
+                if ref_yr2[1]<data2.index.year.min():
+                    ref_yr2[0] = data2.index.year.min()
+                    ref_yr2[1] = ref_yr2[0]+10
+                if ref_yr2[0]<data2.index.year.min():
+                    ref_yr2[0] = data2.index.year.min()
+                    if ref_yr2[1] - ref_yr2[0] < 10:
+                        ref_yr2[1] = ref_yr2[0] + 10
+
+                fig = yrplot_hm(data2, ref_yr2, ag)
+                plt.title(stat2 + ' ' + navi_var + ' anomaly to ' + str(ref_yr2[0]) + '-' + str(ref_yr2[1]))
+                st.pyplot(fig)
+
+        # Re-implement this as a application wide service
+        # expl_md = read_markdown_file('explainer/stripes_m.md')
+        # st.markdown(expl_md, unsafe_allow_html=True)
+
+
+def weather_explorer(config: Config, dataManager: DataManager):
+    """
+    TODO: refactor this whole app into the main_app
+    """
+    # update session state with current data settings
+    data_select(dataManager, config, container=st)
+
+    # check config
+    if config['include_climate']:
+        fig = plt_map(dataManager, sel=config['selected_station'], cm='CORDEX')
+        st.sidebar.plotly_chart(fig)
+        st.sidebar.markdown(
+            '''Map with available stations (<span style="color:blue">blue dots</span>) and selected reference station (<span style="color:magenta">magenta highlight</span>). The climate model grid is given in <span style="color:orange">orange</span> with the selected references as filled dots.''',
+            unsafe_allow_html=True)
+    else:
+        fig = plt_map(dataManager, sel=config['selected_station'])
+        st.sidebar.plotly_chart(fig)
+        st.sidebar.markdown(
+            '''Map with available stations (<span style="color:blue">blue dots</span>) and selected reference station (<span style="color:magenta">magenta highlight</span>).''',
+            unsafe_allow_html=True)
 
+    # switch the topic
+    topic = config['current_topic']
+    if topic == 'Warming':
+        warming_data_plotter(dataManager, config)
+    
+    elif topic == 'Weather Indices':
+        climate_indices(dataManager, config)
 
-            else:
-                yrplot_hm(wdata,ref_yr,ag)
-                plt.title(stat1 + ' ' + navi_var + ' anomaly to ' + str(ref_yr[0]) + '-' + str(ref_yr[1]))
-                st.pyplot()
-
-                sndstat = st.checkbox('Compare to a second station?')
-
-                if sndstat:
-                    stat2 = st.selectbox('Select second station:', [x for x in statios if x != stat1])
-                    data2 = weather[stat2].sel(vars=vari).resample(time='1M').apply(afu).to_dataframe()[stat2]
-                    data2 = data2[~np.isnan(data2)]
-
-                    ref_yr2 = list(ref_yr)
-                    if ref_yr2[1]<data2.index.year.min():
-                        ref_yr2[0] = data2.index.year.min()
-                        ref_yr2[1] = ref_yr2[0]+10
-                    if ref_yr2[0]<data2.index.year.min():
-                        ref_yr2[0] = data2.index.year.min()
-                        if ref_yr2[1] - ref_yr2[0] < 10:
-                            ref_yr2[1] = ref_yr2[0] + 10
-
-                    yrplot_hm(data2, ref_yr2, ag)
-                    plt.title(stat2 + ' ' + navi_var + ' anomaly to ' + str(ref_yr2[0]) + '-' + str(ref_yr2[1]))
-                    st.pyplot()
-
-            # Re-implement this as a application wide service
-            # expl_md = read_markdown_file('explainer/stripes_m.md')
-            # st.markdown(expl_md, unsafe_allow_html=True)
-
-    elif w_topic == 'Weather Indices':
-        climate_indices(stat1,cliproj)
-
-    if cliproj:
+    if config['include_climate']:
         st.markdown(
             '''RCPs are scenarios about possible greenhouse gas concentrations by the year 2100. RCP2.6 is a world in which little further greenhouse gasses are emitted -- similar to the Paris climate agreement from 2015. RCP8.5 was intendent to explore a rather risky, more worst-case future with further increased emissions. RCP4.5 is one candidate of a more moderate greenhouse gas projection, which might be more likely to resemble a realistic situation. It is important to note that the very limited differentiation between RCP scenarios have been under debate for several years. One outcome is the definition of Shared Socioeconomic Pathways (SSPs) for which today, however, there are not very many model runs awailable. For more information, please check with the [Climatescenario Primer](https://climatescenarios.org/primer/), [CarbonBrief](https://www.carbonbrief.org/explainer-how-shared-socioeconomic-pathways-explore-future-climate-change) and this [NatureComment](https://www.nature.com/articles/d41586-020-00177-3)''',
             unsafe_allow_html=True)
 
 
-def main_app():
+def main_app(**kwargs):
+    """Describe the params in kwargs here
+    """
+    # build the config and dataManager from kwargs
+    url_params = st.experimental_get_query_params()
+    config, dataManager = build_config(url_params=url_params, **kwargs)
+
+    st.set_page_config(page_title='Weather Explorer', layout=config.layout)
+
+    # config expander
+    exp = st.sidebar.expander('Data select and config', expanded=False)
+
+    # build the app
     st.header('Weather Data Explorer')
     st.markdown('''In this section we provide visualisations to explore changes in observed weather data. Based on different variables and climate indices it is possible to investigate how climate change manifests itself in different variables, at different stations and with different temporal aggregation.''',unsafe_allow_html=True)
 
-    # TODO: refactor this
-    topics = ['Warming', 'Weather Indices', 'Drought/Flood', 'Agriculture', 'Extreme Events', 'Wind Energy']
-    
     # topic selector
-    topic = components.topic_selector(topic_list=topics, container=st.sidebar)
+    components.topic_selector(config=config, container=st.sidebar, config_expander=exp)
     
     # TODO refactor this
-    weather_explorer(topic)
+    weather_explorer(config, dataManager)
 
 
 if __name__ == '__main__':
-    main_app()
+    import fire
+    fire.Fire(main_app)
diff --git a/ruins/components/topic_select.py b/ruins/components/topic_select.py
index 8289115..4d7b5b9 100644
--- a/ruins/components/topic_select.py
+++ b/ruins/components/topic_select.py
@@ -6,43 +6,43 @@
 from typing import List
 import streamlit as st
 
+from ruins.core import Config
 
-def topic_selector(topic_list: List[str], force_topic_select: bool = True, container=st, **kwargs) -> str:
+
+def topic_selector(config: Config, container=st, config_expander=st) -> str:
     """
-    Select a topic from a list of topics. The selected topic is returned and
-    additionally published to the session cache.
-
-    Parameters
-    ----------
-    topic_list : List[str]
-        List of topics to select from.
-    force_topic_select : bool
-        If False, the dropdown will not be shown if a topic was already
-        selected and is present in the streamlit session cache.
-        Default: True
-    container : streamlit.st.container
-        Container to use for the dropdown. Defaults to main streamlit.
-    **kwargs
-        These keyword arguments are only accepted to directly inject
-        :class:`Config <ruins.config.Config>` objects.
+    TODO: Alex will das dokumentieren....
     """
-    # check if a topic is already present
-    if kwargs.get('no_cache', False):
-        current_topic = kwargs.get('current_topic', None)
-    else:  # pragma: no cover
-        current_topic = st.session_state.get('topic', kwargs.get('current_topic', None))
-    if current_topic is not None and not force_topic_select:
-        return current_topic
+    current_topic = config.get('current_topic')
+    
+    # get topic list
+    topic_list = config['topic_list']
+
+    # get the policy
+    policy = config.get_control_policy('topic_selector')
+
+    # create the control
+    if current_topic is not None:
+        topic = container.selectbox('Select a topic', topic_list)
     
-    # otherwise print select
-    topic = st.selectbox(
-        'Select a topic',
-        topic_list, 
-        index=0 if current_topic is None else topic_list.index(current_topic)
-    )
-
-    # store topic in session cache
-    if current_topic != topic and not kwargs.get('no_cache', False): # pragma: no cover
-        st.session_state['topic'] = topic
+    elif policy == 'show': # pragma: no cover
+        topic = container.selectbox(
+            'Select a topic',
+            topic_list,
+            #index=topic_list.index(config['current_topic'])
+        )
     
-    return topic
+    elif policy == 'hide': # pragma: no cover
+        topic = config_expander.selectbox(
+            'Select a topic',
+            topic_list,
+            #index=topic_list.index(config['current_topic'])
+        )
+    
+    else:
+        topic = current_topic
+
+    # set the new topic
+    if current_topic != topic:
+        st.session_state.current_topic = topic
+
diff --git a/ruins/core/__init__.py b/ruins/core/__init__.py
index d1db6a5..ec72ca5 100644
--- a/ruins/core/__init__.py
+++ b/ruins/core/__init__.py
@@ -1,2 +1,3 @@
 from .config import Config
-from .data_manager import DataManager
\ No newline at end of file
+from .data_manager import DataManager
+from .build import build_config
\ No newline at end of file
diff --git a/ruins/core/build.py b/ruins/core/build.py
index 6247315..62eef1d 100644
--- a/ruins/core/build.py
+++ b/ruins/core/build.py
@@ -2,16 +2,27 @@
 Build a :class:`Config <ruins.core.Config>` and a 
 :class:`DataManager <ruins.core.DataManager>` from a kwargs dict.
 """
-from types import Union, Tuple
+from typing import Union, Tuple, Dict, List
+import streamlit as st
 
 from .config import Config
 from .data_manager import DataManager
 
 
+st.experimental_singleton
+def contextualized_data_manager(**kwargs) -> DataManager:
+    return DataManager(**kwargs)
 
-def build_config(omit_dataManager: bool = False, **kwargs) -> Tuple[Config, Union[None, DataManager]]:
+
+def build_config(omit_dataManager: bool = False, url_params: Dict[str, List[str]] = {}, **kwargs) -> Tuple[Config, Union[None, DataManager]]:
     """
     """
+    # prepare the url params, if any
+    # url params are always a list: https://docs.streamlit.io/library/api-reference/utilities/st.experimental_get_query_params
+    # TODO: This should be sanitzed to avoid injection attacks!
+    ukwargs = {k: v[0] if len(v) == 1 else v for k, v in url_params.items()}
+    kwargs.update(ukwargs)
+
     # extract the DataManager, if it was already instantiated
     if 'dataManager' in kwargs:
         dataManager = kwargs.pop('dataManager')
@@ -22,8 +33,8 @@ def build_config(omit_dataManager: bool = False, **kwargs) -> Tuple[Config, Unio
     config = Config(**kwargs)
 
     if omit_dataManager:
-        return config
+        return config,  None
     else:
         if dataManager is None:
-            dataManager = DataManager(**config)
+            dataManager = contextualized_data_manager(**config)
         return config, dataManager
diff --git a/ruins/core/cache.py b/ruins/core/cache.py
new file mode 100644
index 0000000..50e14d0
--- /dev/null
+++ b/ruins/core/cache.py
@@ -0,0 +1,33 @@
+from typing import Callable, List
+from functools import wraps
+import hashlib
+
+
+LOCAL = dict()
+
+def _hashargs(fname, argnames):
+    h = f'{fname};' + ','.join(argnames)
+    digest = hashlib.sha256(h.encode()).hexdigest()
+
+    return digest
+
+def partial_memoize(hash_names: List[str], store: str = 'local'):
+    def func_decorator(f: Callable):
+        @wraps(f)
+        def wrapper(*args, **kwargs):
+            argnames = [str(a) for a in args if a in hash_names]
+            argnames.extend([str(v) for k, v in kwargs.items() if k in hash_names])
+            
+            # get the parameter hash
+            h = _hashargs(f.__name__, argnames)
+            
+            # check if result exists
+            if h in LOCAL:
+                return LOCAL.get(h)
+            else:
+                # process
+                result = f(*args, **kwargs)
+                LOCAL[h] = result
+                return result
+        return wrapper
+    return func_decorator
diff --git a/ruins/core/config.py b/ruins/core/config.py
index d9d412c..00d6157 100644
--- a/ruins/core/config.py
+++ b/ruins/core/config.py
@@ -1,8 +1,16 @@
+from streamlit import session_state
+import streamlit as st
+
 import os
 from os.path import join as pjoin
 import json
 from collections.abc import Mapping
 
+
+# check if streamlit is running
+if not st._is_running_with_streamlit:
+    session_state = dict()
+
 class Config(Mapping):
     """
     Streamlit app Config object.
@@ -29,6 +37,7 @@ def __init__(self, path: str = None, **kwargs) -> None:
         # path 
         self.basepath = os.path.abspath(pjoin(os.path.dirname(__file__), '..', '..'))
         self.datapath = pjoin(self.basepath, 'data')
+        self.hot_load = kwargs.get('hot_load', False)
 
         # mime readers
         self.default_sources = {
@@ -45,8 +54,14 @@ def __init__(self, path: str = None, **kwargs) -> None:
         }
         self.sources_args.update(kwargs.get('include_args', {}))
 
+        # app management
+        self.layout = 'centered'
+
+        # app content
+        self.topic_list = ['Warming', 'Weather Indices', 'Drought/Flood', 'Agriculture', 'Extreme Events', 'Wind Energy']
+
         # store the keys
-        self._keys = ['debug', 'basepath', 'datapath', 'default_sources', 'sources_args']
+        self._keys = ['debug', 'basepath', 'datapath', 'hot_load', 'default_sources', 'sources_args', 'layout', 'topic_list']
 
         # check if a path was provided
         conf_args = self.from_json(path) if path else {}
@@ -71,8 +86,35 @@ def _update(self, new_settings: dict) -> None:
             if k not in self._keys:
                 self._keys.append(k)
     
+    def get_control_policy(self, control_name: str) -> str:
+        """
+        Get the control policy for the given control name.
+
+        allowed policies are:
+            - show: always show the control on the main container
+            - hide: hide the control on the main container, but move to the expander
+            - ignore: don't show anything
+
+        """
+        if self.has_key(f'{control_name}_policy'):
+            return self.get(f'{control_name}_policy')
+        elif self.has_key('controls_policy'):
+            return self.get('controls_policy')
+        else:
+            # TODO: discuss with conrad to change this
+            return 'show'
+
+    
     def get(self, key: str, default = None):
-        return getattr(self, key, default)
+        if hasattr(self, key):
+            return getattr(self, key)
+        elif key in session_state:
+            return session_state[key]
+        else:
+            return default
+    
+    def has_key(self, key) -> bool:
+        return hasattr(self, key) or hasattr(session_state, key) or key in session_state
     
     def __len__(self) -> int:
         return len(self._keys)
@@ -82,4 +124,14 @@ def __iter__(self):
             yield k
     
     def __getitem__(self, key: str):
-        return getattr(self, key)
+        if hasattr(self, key):
+            return getattr(self, key)
+        elif key in session_state:
+            return session_state[key]
+        else:
+            raise KeyError(f"Key {key} not found")
+    
+    def __setitem__(self, key: str, value):
+        setattr(self, key, value)
+        if key not in self._keys:
+            self._keys.append(key)
diff --git a/ruins/core/data_manager.py b/ruins/core/data_manager.py
index 5f95e3f..2c686d4 100644
--- a/ruins/core/data_manager.py
+++ b/ruins/core/data_manager.py
@@ -67,10 +67,15 @@ class FileSource(DataSource, abc.ABC):
     Abstract base class for file sources. This provides the common interface
     for every data source that is based on a file.
     """
-    def __init__(self, path: str, cache: bool = True, **kwargs):
+    def __init__(self, path: str, cache: bool = True, hot_load = False, **kwargs):
         super().__init__(**kwargs)
         self.path = path
         self.cache = cache
+        
+        # check if the dataset should be pre-loaded
+        if hot_load:
+            self.cache = True
+            self.data = self._load_source()
 
     @abc.abstractmethod
     def _load_source(self):
@@ -94,8 +99,11 @@ class HDF5Source(FileSource):
     """
     HDF5 file sources. This class is used to load HDF5 files.
     """
-    def _load_source(self):
+    def _load_source(self) -> xr.Dataset:
         return xr.open_dataset(self.path)
+    
+    def read(self) -> xr.Dataset:
+        return super(HDF5Source, self).read()
 
 
 class CSVSource(FileSource):
@@ -139,7 +147,7 @@ class DataManager(Mapping):
         The include_mimes can be overwritten by passing filenames directly.
 
     """
-    def __init__(self, datapath: str = None, cache: bool = True, debug: bool = False, **kwargs) -> None:
+    def __init__(self, datapath: str = None, cache: bool = True, hot_load = False, debug: bool = False, **kwargs) -> None:
         """
         You can pass in a Config as kwargs.
         """
@@ -148,9 +156,9 @@ def __init__(self, datapath: str = None, cache: bool = True, debug: bool = False
             from ruins.core import Config
             self.from_config(**Config(**kwargs))
         else:
-            self.from_config(datapath=datapath, cache=cache, debug=debug, **kwargs)
+            self.from_config(datapath=datapath, cache=cache, hot_load=hot_load, debug=debug, **kwargs)
     
-    def from_config(self, datapath: str = None, cache: bool = True, debug: bool = False, **kwargs) -> None:
+    def from_config(self, datapath: str = None, cache: bool = True, hot_load: bool = False, debug: bool = False, **kwargs) -> None:
         """
         Initialize the DataManager from a :class:`Config <ruins.core.Config>` object.
         """
@@ -158,6 +166,7 @@ def from_config(self, datapath: str = None, cache: bool = True, debug: bool = Fa
         self._config = kwargs
         self._datapath = datapath
         self.cache = cache
+        self.hot_load = hot_load
         self.debug = debug
 
         # file settings
@@ -225,7 +234,7 @@ def add_source(self, path: str, not_exists: str = 'raise') -> None:
             
             # add the source
 #            args = self._config.get(basename, {})
-            args.update({'path': path, 'cache': self.cache})
+            args.update({'path': path, 'cache': self.cache, 'hot_load': self.hot_load})
             self._data_sources[basename] = BaseClass(**args)
         else:
             if not_exists == 'raise':
@@ -255,7 +264,7 @@ def __iter__(self):
         for name in self._data_sources.keys():
             yield name
     
-    def __getitem__(self, key: str):
+    def __getitem__(self, key: str) -> DataSource:
         """Return the requested datasource"""
         return self._data_sources[key]
 
diff --git a/ruins/plotting/kde.py b/ruins/plotting/kde.py
index 18ad7fb..ff96635 100644
--- a/ruins/plotting/kde.py
+++ b/ruins/plotting/kde.py
@@ -122,4 +122,4 @@ def kde(data, cmdata='none', split_ts=1, cplot=True, eq_period=True):
 
     ax.set_ylabel('Occurrence (KDE)')
 
-    return ax
+    return fig, ax
diff --git a/ruins/plotting/maps.py b/ruins/plotting/maps.py
index cdbcd1a..9a18df3 100644
--- a/ruins/plotting/maps.py
+++ b/ruins/plotting/maps.py
@@ -1,23 +1,29 @@
 import streamlit as st
 import plotly.graph_objs as go
 import plotly.express as px
+import pandas as pd
 import numpy as np
 
+from ruins.core import DataManager
+from ruins.core.cache import partial_memoize
+
+
+@partial_memoize(hash_names=['sel', 'cm'])
+def plt_map(dataManager: DataManager, sel='all', cm='none') -> go.Figure:
+    # cordex_grid = xr.open_dataset('data/CORDEXgrid.nc')
+    # cimp_grid = xr.open_dataset('data/CMIP5grid.nc')
+    # stats = pd.read_csv('data/stats.csv', index_col=0)
+    cordex_grid = dataManager['CORDEXgrid'].read()
+    cimp_grid = dataManager['CMIP5grid'].read()
+    stats = dataManager['stats'].read()
 
-def plt_map(sel='all',cm='none'):
-    # TODO remove this part
-    import xarray as xr
-    import pandas as pd
-    dummy = xr.open_dataset('data/CORDEXgrid.nc')
-    dummy5 = xr.open_dataset('data/CMIP5grid.nc')
-    stats = pd.read_csv('data/stats.csv', index_col=0)
     stats['ms'] = 15.
     stats['color'] = 'gray'
 
     mapbox_access_token = 'pk.eyJ1IjoiY29qYWNrIiwiYSI6IkRTNjV1T2MifQ.EWzL4Qk-VvQoaeJBfE6VSA'
     px.set_mapbox_access_token(mapbox_access_token)
 
-    nodexy = pd.DataFrame([dummy.lon.values.ravel(), dummy.lat.values.ravel()]).T
+    nodexy = pd.DataFrame([cordex_grid.lon.values.ravel(), cordex_grid.lat.values.ravel()]).T
     nodexy.columns = ['lon', 'lat']
     nodexy['hov'] = 'CORDEX grid'
 
@@ -127,8 +133,8 @@ def add_cmpx(cm):
                             [13, 13], [14, 13], [15, 13], [16, 13], [17, 13], [18, 13], [19, 14], [20, 14], [21, 13]]
         for cc in maskcordex_coast:
             fig.add_trace(go.Scattermapbox(
-                lat=[dummy.lat.values[tuple(cc)]],
-                lon=[dummy.lon.values[tuple(cc)]],
+                lat=[cordex_grid.lat.values[tuple(cc)]],
+                lon=[cordex_grid.lon.values[tuple(cc)]],
                 mode='markers',
                 marker=go.scattermapbox.Marker(
                     size=8,
@@ -143,11 +149,11 @@ def add_cmpx(cm):
     #fig = px.scatter_mapbox(nodexy, lat='lat', lon='lon', center={'lat': 53.0, 'lon': 8.3}, zoom=5, opacity=0.1, hover_data=['hov'])
     fig = px.scatter_mapbox(stats, lat='lat', lon='lon', center={'lat': 53.0, 'lon': 8.6}, zoom=5, size='ms', opacity=0.8, color='color', hover_data=['Station name', 'lat', 'lon'], size_max=10)
     if cm != 'none':
-        fig = lin_grid(fig, dummy)
-        fig = lin_grid(fig, dummy5, '#2c7fb8')
+        fig = lin_grid(fig, cordex_grid)
+        fig = lin_grid(fig, cimp_grid, '#2c7fb8')
         fig = add_cmpx(cm)
     fig = add_stats(sel)
     fig.update_layout(showlegend=False,width=300, height=350,margin=dict(l=10, r=10, b=10, t=10))  # ,center={'lat':54.0,'lon':8.3}, zoom=7)
-    st.sidebar.plotly_chart(fig)
-    return
+    # st.sidebar.plotly_chart(fig)
+    return fig
 
diff --git a/ruins/plotting/weather_data.py b/ruins/plotting/weather_data.py
index 5a52116..77cfea0 100644
--- a/ruins/plotting/weather_data.py
+++ b/ruins/plotting/weather_data.py
@@ -1,6 +1,7 @@
 import numpy as np
 import pandas as pd
 import matplotlib.pyplot as plt
+import seaborn as sns
 
 
 def yrplot_hm(sr, ref=[1980, 2000], ag='sum', qa=0.95, cbar_title='Temperature anomaly (K)', cmx='coolwarm', cmxeq=True, li=False):
@@ -40,7 +41,7 @@ def yrplot_hm(sr, ref=[1980, 2000], ag='sum', qa=0.95, cbar_title='Temperature a
     if ag == 'sum':
         dummy.iloc[:, 13] = dummy.iloc[:, 13] / 12
 
-    plt.figure(figsize=(8,len(dummy)/15.))
+    fig = plt.figure(figsize=(8,len(dummy)/15.))
     ax = sns.heatmap(dummy, cmap=cmx, vmin=vxL, vmax=vxU, cbar_kws={'label': cbar_title})
 
     if ref == None:
@@ -65,7 +66,8 @@ def yrplot_hm(sr, ref=[1980, 2000], ag='sum', qa=0.95, cbar_title='Temperature a
 
     ax.set_ylabel('Year')
     ax.set_xlabel('Month          ')
-    return
+    
+    return fig
 
 
 def monthlyx(dy, dyx=1, ylab='T (°C)', clab1='Monthly Mean in Year', clab2='Monthly Max in Year', pls='cividis_r'):
diff --git a/ruins/tests/test_config.py b/ruins/tests/test_config.py
index 548047b..75deb4d 100644
--- a/ruins/tests/test_config.py
+++ b/ruins/tests/test_config.py
@@ -3,6 +3,7 @@
 """
 import os
 import json
+import pytest
 
 from ruins import core
 
@@ -62,3 +63,11 @@ def test_config_as_dict():
     # check default get behavior
     assert c.get('doesNotExist') is None
     assert c.get('doesNotExists', 'foobar') == 'foobar'
+
+
+def test_config_key_error():
+    """An unknown key should throw a key error"""
+    c = core.Config()
+
+    with pytest.raises(KeyError):
+        c['doesNotExist']
diff --git a/ruins/tests/test_topic_selector.py b/ruins/tests/test_topic_selector.py
deleted file mode 100644
index 0a091c8..0000000
--- a/ruins/tests/test_topic_selector.py
+++ /dev/null
@@ -1,23 +0,0 @@
-from ruins import components
-
-
-def test_default():
-    """Test default behavior"""
-    assert components.topic_selector(['a', 'b', 'c'], no_cache=True) == 'a'
-
-
-def test_current_preset():
-    """Test with current topic set"""
-    topic = components.topic_selector(['a', 'b', 'c'], current_topic='b', no_cache=True)
-
-    assert topic == 'b'
-
-
-def test_no_force_render():
-    """Test with rendering disabled"""
-    # default
-    assert components.topic_selector(['a', 'b', 'c'], force_topic_select=False, no_cache=True) == 'a'
-
-    # with current topic set
-    assert components.topic_selector(['a', 'b', 'c'], current_topic='b', force_topic_select=False, no_cache=True) == 'b'
-    
\ No newline at end of file
diff --git a/ruins/tests/test_weather.py b/ruins/tests/test_weather.py
index 64fe7c7..968c003 100644
--- a/ruins/tests/test_weather.py
+++ b/ruins/tests/test_weather.py
@@ -1,27 +1,20 @@
 from ruins.apps import weather
 from ruins.tests.util import get_test_config
 
-from ruins.core import DataManager
+from ruins.core import DataManager, Config
 
-
-# TODO use the config and inject the dedub config here
-
-# TODO only run this test when the Value Error is solved
-#def test_run_app():
-#    """Make sure the appp runs without failing"""
-#    weather.main_app()
+# create the test config
+config = get_test_config()
 
 
 def test_climate_indices():
     """Test only climate indices """
-    conf = get_test_config()
-    dm = DataManager(**conf)
-
-    w = dm['weather'].read()
-    c = dm['cordex_coast'].read()
+    # add the include_climate config
+    config['include_climate'] = True
+    dm = DataManager(**config)
 
     # run
-    weather.climate_indices(w, c)
+    weather.climate_indices(dataManager=dm, config=config)
 
 
 def test_climate_indi():