data-for-good-vancouver/brightside-datathon

Code repository for the Brightside Datathon analyses

Data for Good/Brightside Datathon

Table of Contents

• Project Information
  • Dependencies
• Installation
• Objectives

---

Project Information

This is the repository for the Brightside/Data for Good 2021 Datathon.

Dependencies

Currently, git is required to have access to this repository.

---

Getting Started

Welcome to the Github repository for the August 14, 2021 Brightside/Data for Good Datathon. Create a folder and upload your analysis here with relevant unique identifiers (e.g a team name) to identify yourself. Make sure the README contains the names of the team members. Include a description of the dependencies of your analyses. If possible, make sure the analysis is reproducible (e.g Jupyter Notebook, Snakemake, etc).

Installation

To clone the repository, run the following shell command:

git clone 

Objectives

The objective of this event is to help Brightside homes to make interesting observations from their data. The datathon will be open-ended and you can take the analysis wherever you would like. Our partners are looking for us to identify patterns and themes in the data that can build an accurate story/narrative about the experience of the people they work with and support.

For example:
What are residents seeing/experiencing?
Are there food security issues?
Are there changes over the years? Can we quantify any of these changes?
Are certain populations experiencing different things than others?
Are there geographic variances in the data?

From a slightly more perfunctory lens - these efforts can be broadly split into a few categories.

1. Descriptive statistics regarding certain variables of interests or other exploratory data analysis
2. Observing trends in the collected data or uncovering underlying relationships using linear models
3. Integrating the collected housing data with data from external databases (ex. food bank, crimedata, etc.)
4. Coming up with interesting/innovative and informative ways to visualize these observations (e.g heatmaps, PCA, tSNE)
5. Predictive modelling (For example: Using machine learning via sklearn's built-in random forest, SVM, etc modules)

Again, this datathon is intended to be open-ended and we want to encourage those examining the data to be creative. The questions we have posed above are broad and are intended to be a start. You, the analyst(s), chooses where to go based on what you are interested in and perhaps what you think will be most useful.

---