Getting Started

Waste and Resources

Our client stated that the goal was to have report materials that can make for a more educated "garbage geek culture". She would like the visualizations to update

Technology

The goal of this project from a technology perspective was to show the power of the open source python ecosystem to address the data cataloging, calculation of key features, and interactive visualization needs of our client. This pythonic approach is equally applicable in other contexts such as banks where the data resources are located in seperate databases and complex bussiness logic is required to produce interactive visualizations.

Data Processing and Data Catalog (Intake)

The waste and resources department has their data in one large excel file containing multiple sheets with several matrices each for different types of waste. We processed this excel sheet using pandas, saving the data out into one csv file per matrix. A data catalog was used to unify the data as a single resource and expose the data as a dataframe. Allowing each datasource to be manipulated in the same way. Additionally it allow the data to update with each new month of data automatically.

import intake

catalog = intake.open_catalog('catalog.yml')
catalog.solid_waste_commodity_recycling

Dashboard Visualization (Bokeh + LeafletJS)

Our client wanted to have accessible infographics about the current waste month by month, plotting, and resources to create a more educated "garbage geek culture". We thought a dynamic single page dashboard application would address the needs of our customer. Uses a bokeh python server running in background to provide richer interactive experience.

Filters

visitors can apply filters which update all graphs:

• date range of data
• change units that all plots are being calculated with

Dynamic "Did you know questions?"

Did you know questions python generated, randomly selected, and update with current monthly data.

Dynamic Text Fields

Autoupdating text fields within dashboard that are calculated from data.

Development

1. Install anaconda version doesn't matter

2. Import anaconda environment: conda env create -n hackathon -f hackathon.yml

3. Switch to the new environment: source activate hackathon

4. Add the environment to your jupyter: python -m ipykernel install --user --name hackathon --display-name "Hackathon".

Deployment

bokeh serve --show visualization