Name: sedesol-public
Owner: Data Science for Social Good
Description: null
Created: 2016-08-24 20:23:09.0
Updated: 2017-01-17 21:55:09.0
Pushed: 2017-01-17 22:05:49.0
Homepage: null
Size: 10569
Language: Python
GitHub Committers
| User | Most Recent Commit | # Commits |
|---|
Other Committers
| User | Most Recent Commit | # Commits |
|---|
Project name: Enhancing the Distribution of Social Services in Mexico
Partner: SEDESOL
Issues: https://github.com/dssg/sedesol/issues
The Ministry for Social Development (SEDESOL) operates a range of social service programs to fight poverty in Mexico. One of their major challenges is how to effectively identify and target the families in most need. DSSG is working with SEDESOL to develop data-driven methods that can identify: (i) the socio-economic needs (food, health, education, quality of dwellings, basic housing services, and social security) of eligible families in poverty, and (ii) people who might be under-reporting their socio-economic conditions in order to qualify for programs. We hope that these targeting and detection strategies can improve the resource allocation and the targeting of programs towards those in most need. Read more about DSSG and our project here.
Our Technical Plan, which contains the detailed explanation about the project
as well as results and future work, can be found on the docs/technical-plan
folder.
This project uses pyenv. We prefer
pyenv to virtualenv since it is simpler without sacrificing any useful
feature.
Following the standard practice in pyenv the version of python is specified
in .python-version. In this particular case is 3.5.2.
These can be found on the requirements.txt file.
To ease the setup, a Dockerfile is provided which builds an image with all
dependencies included and properly configured. Several Docker images are setup
with Docker compose. For further information see the
README
in infrastructure.
Please take into account that this process downloads and installs all dependencies.
For information on how to setup Docker, see the official docs.
Once you have set up the environment, you can start using the pipeline. There are two different pipelines which can be run either individually or jointly, pub-imputation and underreporting. The general process of both pipelines is:
Run the following
git clone https://github.com/dssg/sedesol.gitcd sedesolmake preparemake deployThe pipeline assumes the existence of all raw data already in Postgres tables,
inside a raw schema. Instructions and scripts for transforming and uploading
both the data SEDESOL provided and the publicly available data can be found at
the /etl folder, specifically on the db_ingestion* files.
aws s3 and use smart_open for a transparent
write/read operations