tauseef1234 / Data-Visualization

Interactive Bokeh Dashboard - Using Google cloud public dataset

Chicago crime data - Interactive Dashboard

The two interactive dashboards developed using open source visualization library BOKEH are aimed to derive key insights in criminal offences committed in the city of Chicago for the year 2019. The dashboard shows a choropleth map of the city of Chicago that has counties color coded by the offence count, a datatable and a horizontal bar chart of the offence types.

Input Data

The source data for this dashboard is extracted from Google cloud public dataset

Data Extraction - Python library bigquery was used to extract data from Google cloud. The steps to create Google API credentials can be figured out using Google authentication webpage. The following code snippet was run to fetch the data.

pip install --upgrade google-cloud-bigquery        # Install library
os.environ["GOOGLE_APPLICATION_CREDENTIALS"]=r"C:\Users\*******.json"   # Replace the file name by your own google API credentials
client = bigquery.Client()          # Start the BigQuery Client
QUERY = ('SELECT * FROM `bigquery-public-data.chicago_crime.crime`')              # Input Query Syntax
query_job = client.query(QUERY)     # Start Query API Request
query_result = query_job.result()   # Get Query Result
df = query_result.to_dataframe()    # Save the Query to Dataframe

The pickle file chicago_crime.pkl is generated from the df dataframe using the above code snippet

Google cloud stores data from 2001 till 2019 but this dashboard uses data for year 2019 only.

The Boundaries.geojson (mapping data) json file is extracted from the Chicago city website

Dashboard 1

Dashboard1 Link

This dashboard (chicago3.py) has three key components:

1. Choropleth map - A geomap in which county areas are color coded by the count of offence in the city of Chicago for the year 2019.
2. Datatable - A datatable view of the count by offence type
3. Bar chart - A horizontal bar chart is used to display the count by offence type

How the dashboard works

When clicked on a county on the geomap, the datatable and horizontal bar chart get refreshed with the offence count by that county. When clicked on an offence type on datatable, the geomap gets resfreshed with the offence count by the offence type clicked on the data table. The dashboard is developed using Python interactive data visualization library bokeh.

Dashboard 2

Dashboard2 Link

The second dashobard (chicago_2019.py) deep dives into the distribution of offences by month and offence type. The key components of this dashboard are:

1. Choropleth map - A geomap in which county areas are color coded by the count of offence in the city of Chicago for the year 2019.
2. Datatable - A datatable view of the percentage of offence by day of the week and time of the day
3. Bar chart - A horizontal bar chart is used to display the count by offence type and month

References:

1. Bokeh documentation is the best place to get started with the library - great resource
2. This great blog from Jim King on creating an interactive geographic map using bokeh