Create app.py

app.py

@@ -0,0 +1,247 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+import pandas as pd 
+import geopandas as gpd
+import matplotlib.pyplot as plt
+import streamlit as st
+import plotly.express as px
+import plotly.graph_objects as go
+
+# read shapefile 
+gdf = gpd.read_file(r'C:\Users\hp\Downloads\mys_admb_unhcr_20210211_shp\mys_admbnda_adm2_unhcr_20210211.shp')
+
+
+# plot with colors, and right dimensions
+gdf.plot(column='ADM2_EN', cmap=None, legend=None, figsize=(20, 20))
+
+# gdf[["Shape_Leng", "Shape_Area", "ADM2_EN", "ADM2_PCODE", "ADM2_REF", "ADM2ALT1EN", "ADM2ALT2EN", "ADM1_EN", "ADM1_PCODE", "ADM0_EN", "ADM0_PCODE", "date", "validOn", "validTo", "geometry"]]
+newgdf = gdf[["ADM2_EN", "ADM1_EN", "Shape_Leng", "Shape_Area", "ADM2_PCODE", "ADM1_PCODE",
+        "date", "validOn","geometry"]]
+
+newgdf.columns=["District", "State", "Shape_Leng", "Shape_Area", "District Postcode", "State Postcode", "date", "validOn","geometry"]
+newgdf.plot(column='District', cmap=None, legend=None, figsize=(20, 20))
+
+poi = gpd.read_file(r'C:\Users\hp\Downloads\hotosm_mys_points_of_interest_points_shp/hotosm_mys_points_of_interest_points.shp')
+simple_poi = poi[["osm_id", "man_made", "geometry"]].copy()
+
+@st.cache(suppress_st_warning=True)
+def fit_district_coord_to_state():
+    st.write("Cache miss: fit_district_coord_to_state() ran") 
+    poi_temp = gpd.read_file(r'C:\Users\hp\Downloads\hotosm_mys_points_of_interest_points_shp/hotosm_mys_points_of_interest_points.shp')
+    gdf_temp = gpd.read_file(r'C:\Users\hp\Downloads\mys_admb_unhcr_20210211_shp\mys_admbnda_adm2_unhcr_20210211.shp')
+    test_state = []
+    test_dist = []
+    k = 1
+    for i in poi_temp["geometry"]:
+        j = 0
+        while not i.within(gdf_temp.iat[j, 14]):
+            j = j + 1
+            if j >= 144: 
+                break 
+        if j >= 144: 
+            test_dist.append("None")
+            test_state.append("None")
+        else: 
+            test_dist.append(gdf_temp.iat[j, 2])
+            test_state.append(gdf_temp.iat[j, 7]) 
+    return(test_dist, test_state)
+
+temp = fit_district_coord_to_state()
+poi["District"]= temp[0]
+poi["State"] = temp[1]
+
+poi_none = poi[poi.District == "None"]
+kv_gdf = gdf[gdf.ADM1_EN == "Selangor"]
+kv_gdf = kv_gdf.append(gdf[(gdf.ADM1_EN == "W.P. Putrajaya")], ignore_index=True)
+kv_gdf = kv_gdf.append(gdf[(gdf.ADM1_EN == "W.P. Kuala Lumpur")], ignore_index=True)
+kv_gdf = kv_gdf.append(gdf[(gdf.ADM2_EN == "Seremban")], ignore_index=True)
+
+# find missing POI within KV and assign correct District and State 
+# ID = 2083151347, 5928578725 ; to be changed from None to Klang 
+#poi[poi.osm_id == 2083151347] # index of 3848
+#poi[poi.osm_id == 5928578725] # index of 32419
+poi.at[3848, 'State'] = "Selangor"
+poi.at[3848, 'District'] = "Klang"
+poi.at[32419, 'State'] = "Selangor"
+poi.at[32419, 'District'] = "Klang"
+
+kv_poi = poi[poi.State == "Selangor"]
+kv_poi = kv_poi.append(poi[poi.State == "W.P. Putrajaya"], ignore_index=True)
+kv_poi = kv_poi.append(poi[poi.State == "W.P. Kuala Lumpur"], ignore_index=True)
+kv_poi = kv_poi.append(poi[poi.District == "Seremban"], ignore_index=True)
+
+
+
+districts = kv_gdf['ADM2_EN'].unique()
+
+shop_1 = sorted(list(filter(None, kv_poi['shop'].unique())))
+shop_2 = [name.title() for name in shop_1]
+amenity_1 = sorted(list(filter(None, kv_poi['amenity'].unique())))
+amenity_2 = [name.title() for name in amenity_1]
+tourism_1 = sorted(list(filter(None, kv_poi['tourism'].unique())))
+tourism_2 = [name.title() for name in tourism_1]
+man_made_1 = sorted(list(filter(None, kv_poi['man_made'].unique())))
+man_made_2 = [name.title() for name in man_made_1]
+types = {
+    'Shops' : shop_2, 
+    'Amenity': amenity_2,
+    'Tourism': tourism_2,
+    'Man Made': man_made_2
+}
+category = sorted(['Shops', 'Amenity', 'Tourism', 'Man Made'])
+
+MY_HASH = {
+    pd.DataFrame: lambda _: None,
+    str: lambda _: None,
+    int: lambda _: None,
+    list: lambda _: None,
+    gpd.GeoDataFrame: lambda _: None
+}
+@st.cache(suppress_st_warning=True, hash_funcs=MY_HASH)
+def build_df():
+    st.write("build_df() ran")
+    f_kv_poi = kv_poi.copy()
+    f_districts = kv_gdf['ADM2_EN'].unique()
+    f_shopc = []
+    f_amenc = []
+    f_tourc = []
+    f_manmc = []
+    f_coord = []
+    f_shopt = pd.DataFrame()
+    f_ament = pd.DataFrame()
+    f_tourt = pd.DataFrame()
+    f_manmt = pd.DataFrame()
+    for d in f_districts: 
+        summary = f_kv_poi[f_kv_poi.District == str(d)].count()
+        f_shopc.append(summary[11])
+        f_amenc.append(summary[6])
+        f_tourc.append(summary[1])
+        f_manmc.append(summary[13])
+        f_shopt = pd.concat([f_shopt, f_kv_poi[f_kv_poi.District == str(d)].shop.value_counts()], axis=1)
+        f_ament = pd.concat([f_ament, f_kv_poi[f_kv_poi.District == str(d)].amenity.value_counts()], axis=1)
+        f_tourt = pd.concat([f_tourt, f_kv_poi[f_kv_poi.District == str(d)].tourism.value_counts()], axis=1)
+        f_manmt = pd.concat([f_manmt, f_kv_poi[f_kv_poi.District == str(d)].man_made.value_counts()], axis=1)
+    return(f_shopt, f_ament, f_tourt, f_manmt, f_shopc, f_amenc, f_tourc, f_manmc)
+
+temp_store = build_df()
+shopt = temp_store[0]
+ament = temp_store[1]
+tourt = temp_store[2]
+manmt = temp_store[3]
+shopc = temp_store[4]
+amenc = temp_store[5]
+tourc = temp_store[6]
+manmc = temp_store[7]
+
+shopt.columns = districts
+ament.columns = districts 
+tourt.columns = districts
+manmt.columns = districts
+total_df= pd.DataFrame()
+
+total_df = total_df.set_axis(districts, axis=0)
+
+total_df= pd.concat([total_df, shopt.transpose(), ament.transpose(), tourt.transpose(), manmt.transpose()], axis=1)
+total_df['Amenity'] = amenc
+total_df['Man Made'] = manmc
+total_df['Shops'] = shopc
+total_df['Tourism'] = tourc
+total_df = total_df.fillna(0)
+total_df = (100. * total_df / total_df.sum()).round(2)
+temp = pd.DataFrame(kv_gdf.geometry).transpose()
+temp.columns = districts
+total_df = pd.concat([total_df, temp.transpose()], axis=1)
+
+total_df['districts'] = districts
+total_df.columns = [x.title() for x in total_df.columns]
+total_gdf = gpd.GeoDataFrame(total_df, crs="EPSG:4326", geometry='Geometry')
+
+
+st.title ("An Analysis of the Points of Interest (POIs) within Klang Valley ")
+st.header("Definition of POIs")
+st.write('We have four categories, "Amenity", "Man Made", "Shops", and "Tourism". And with each category, there are multiple types of POIs.')
+
+# insert simple description of certain POIs measured. 
+
+st.header("Visualization")
+st.write('Please select the category and type to view the choropleth maps.')
+# #    st.write("Select Category and Type")
+# #    category_option = st.selectbox('Select Category', category)
+# #type_option = st.selectbox('Select Type', types[str(category_option)])
+
+#     # adding "select" as the first and default choice
+#     category_option = st.selectbox('Select Category', options=['select']+list(types.keys()))
+#     # display selectbox 2 if manufacturer is not "select"
+#     if category_option != 'select':
+#         type_option = st.selectbox('Select Type', options=types[category_option])
+#     submitted = st.form_submit_button("Submit")
+#     if submitted:
+#         st.write('You selected ' + category_option + ' ' + type_option)
+
+category_option = st.selectbox('Select Category', options=['Select Category']+list(types.keys()))
+
+if category_option != 'Select Category':
+    max_category = total_gdf.loc[total_gdf[str(category_option)].idxmax()][303]
+    min_category = total_gdf.loc[total_gdf[str(category_option)].idxmin()][303]
+    st.write(f"You have selected Category **{category_option}**")
+    fig_category = go.Figure(px.choropleth_mapbox(total_gdf,
+                   geojson=total_gdf.geometry,
+                   locations=total_gdf.Districts,
+                   color=str(category_option),
+                   center={"lat": 3.140853, "lon": 101.693207},
+                   mapbox_style="carto-positron",
+                   zoom=7.5, 
+                   color_continuous_scale = "blues", 
+                   opacity=0.7,
+                   labels={str(category_option): "%"}
+                   ))
+
+    fig_category.update_geos(fitbounds="locations", visible=False)
+    fig_category.update_layout(title_text=f"Choropleth Map for {category_option} in Klang Valley Districts", 
+        paper_bgcolor="#F0F2F6",
+        margin={"r": 30, "t": 50, "l": 1, "b": 1})
+    st.plotly_chart(fig_category)
+    st.write(f"Max: {max_category}, Min: {min_category}")
+
+    type_option = st.selectbox('Select Type', options=['Select Type']+types[category_option])
+
+    if type_option != 'Select Type': #st.submit('Confirm')
+        st.write(f"You have selected Type **{type_option}** for Category {category_option}.")
+        max_type = total_gdf.loc[total_gdf[str(type_option)].idxmax()][303]
+        min_type = total_gdf.loc[total_gdf[str(type_option)].idxmin()][303]
+
+        hi = total_gdf[[str(type_option), 'Geometry', 'Districts']]
+        fig_type = go.Figure(px.choropleth_mapbox(hi,
+                           geojson=hi.Geometry,
+                           locations=hi.Districts,
+                           color=str(type_option),
+                           center={"lat": 3.140853, "lon": 101.693207},
+                           mapbox_style="carto-positron",
+                           zoom=7.5, 
+                           color_continuous_scale = "bupu", 
+                           opacity=0.7, 
+                           labels={str(type_option): "%"}
+                           ))
+
+        fig_type.update_geos(fitbounds="locations", visible=False)
+        fig_type.update_layout(title_text=f"Choropleth Map for {type_option} ({category_option}) in Klang Valley Districts", 
+            paper_bgcolor="#F0F2F6",
+            margin={"r": 30, "t": 50, "l": 1, "b": 1})
+        st.plotly_chart(fig_type)
+        st.write(f"Max: {max_type}, Min: {min_type}")
+
+
+st.header("Comments")
+st.write('From comparing the district with the most and least respective POI, we can see that Kuala Lumpur is the district with the most POIs.')
+st.write('We can also see that Kuala Selangor has the least POIs in general.')
+st.write('However, we also can see that the information from these maps are not very useful.')
+
+st.header('Improvements')
+st.write('What the visualizations lack is a comparison with other factors. Adding information such as **population density**, **number of roads or rivers**, or other information that would allow a more in depth comparison would result in a better analysis of the POIs.')
+st.write('This would allow us to correlate and understand why certain districts have more or less POIs than others.')
+st.write('Another point of improvement could be allowing user to zoom in a specific district to see the precise locations of the respective POIs.')
+
+st.header("Questions")
+st.write('Which district has the most "Shops"? Which shops are most prevalent in Petaling? Does more people lead to more petrol stations?')
+