Truth Tree Client

Authors: Ben Hurst, Justin Hedani, Greg Schundler, Lance Warrick

Version: 0.1.0

Back-end GitHub Repo: https://github.com/TruthTree/TruthTreeAPI

Front-end Tools: React, Mapbox GL JS, and D3 (VictoryJS)

---

Table of contents

• Overview
• Current Work
• Next Steps
• Getting Started
• Change Log

---

Overview

Truth Tree is an interactive, GIS-based data visualization tool presenting the breakdown of government spending and related metrics in an effort to provide a unified resource to gauge public investment effectiveness.

Our long-term goal is a community-driven platform where registered users can create their own data visualizations, publish stories, compare trends, and engage with the public and policymakers on the issues that matter most.

Key Features:

• React.js components will be configured to consume TruthTreeAPI endpoints. State and County statistics will be dynamically loaded with current data from API call.

• Interactive map rendered using Mapbox GL and a custom GeoJSON modification/generation script (assets/modify_geojson.py).

• Charts enabled by VictoryJS.

Current Work

Phase One efforts consist of converting the Atkinson School of Management's Government Finance Database (State & County datasets for the time being) into a highly accessible and interactive web application to provide a baseline exploration tool.

If you're interested in contributing, we reccomend checking out the following resources:

• Democracy Lab Profile
• Government Finance Database
• IHME Opioid Study (Our Phase One efforts will be based heavily on presenting a similar tool with the GFD dataset)

Other useful references:

• Our Wolrd In Data
• ClearGov
• Gapminder
• IHME Viz Hub

Next Steps

As a preliminary proof of concept, this repo generates a chloropleth map presenting state correctional spending (total, per capita, and as a percent of total expenditures). Aside from general cleanup, we could use much help in:

1. Integrating the front-end to dynamically call (perhaps through a typeahead-driven form) a specified column in the dataset (not just correctional spending) from the backend and render out the associated chloropleth on the frontend.
2. Update the chloropeth layer to show county-level data on zoom (https://www.mapbox.com/mapbox-gl-js/example/updating-choropleth/). This will require making appends to the county geojson file.

Key Blockers:

1. In regards to time series analysis, county level data is only complete in years ending with a 2 or 7 (when a census on government financial statistics is conducted). As a consequene, interim years only contain about half of the ~3000 counties in the country. Reference pg. 3-4 of The Government Finance Database research article
2. The geojson file for counties was compiled in 2010. However, counties periodically get redrawn. Thus, providing a sliding time scale on the map would require accounting for this discrepancy.

---

Getting Started in Development (using yarn and pipenv)

In a terminal instance:

1. git clone https://github.com/TruthTree/ClientTruthTree.git
2. cd ClientTruthTree/
3. yarn install to install required dependencies.
4. touch env and add your Mapbox access key as REACT_APP_MAPBOX_ACCESS_TOKEN.
5. cd src/assets/
6. pipenv shell
7. pipenv install
8. Download State and County datasets from The Government Finance Database and place in assets/ directory.
9. Download 20m State & County geojson files from http://eric.clst.org/tech/usgeojson/ and place in assets/ directory (rename states_2010.json & counties_2010.json)
10. python3 modify_geojson.py
11. exit
12. cd ../../
13. yarn start

Change Log

1-15-2018 11:00am - Initial Scaffolding.

Back to top