Data folder restructuring in preparation for 361 (#376)

* initial checkin

* gitignore and docker-compose update

* readme update and error on hud

* encoding issue

* one more small README change

* data roadmap re-strcuture

* pyproject sort

* small update to score output folders

* checkpoint

* couple of last fixes

.gitignore

@@ -130,15 +130,15 @@ dmypy.json
 cython_debug/
 
 # Ignore dynaconf secret files
-score/.secrets.*
+*/data-pipeline/.secrets.*
 
 # ignore data
-score/data
+*/data-pipeline/data
-score/data/census
+*/data-pipeline/data/census
-score/data/tiles
+*/data-pipeline/data/tiles
-score/data/tmp
+*/data-pipeline/data/tmp
-score/data/dataset
+*/data-pipeline/data/dataset
-score/data/score
+*/data-pipeline/data/score
 
 # node
 node_modules

data/data-pipeline/Dockerfile

@@ -25,7 +25,7 @@ RUN add-apt-repository ppa:ubuntugis/ppa
 RUN apt-get -y install gdal-bin
 
 # Prepare python packages
-WORKDIR /score
+WORKDIR /data-pipeline
 RUN pip3 install --upgrade pip setuptools wheel
 COPY . .
 

data/data-pipeline/README.md

@@ -93,30 +93,30 @@ We use Docker to install the necessary libraries in a container that can be run
 
 To build the docker container the first time, make sure you're in the root directory of the repository and run `docker-compose build`.
 
-After that, to run commands type the following:
+Once completed, run `docker-compose up` and then open a new tab or terminal window, and then run any command for the application using this format:
+`docker exec j40_data_pipeline_1 python3 application.py [command]`
 
-- Get help: `docker run --rm -it j40_score /bin/sh -c "python3 application.py --help"`
+Here's a list of commands:
-- Clean up the census data directories: `docker run --rm -it j40_score /bin/sh -c "python3 application.py census-cleanup"`
-- Clean up the data directories: `docker run --rm -it j40_score /bin/sh -c "python3 application.py data-cleanup"`
-- Generate census data: `docker run --rm -it j40_score /bin/sh -c "python3 application.py census-data-download"`
-- Run all ETL processes: `docker run --rm -it j40_score /bin/sh -c "python3 application.py etl-run"`
-- Generate Score: `docker run --rm -it j40_score /bin/sh -c "python3 application.py score-run"`
 
-## Log visualization
+- Get help: `docker exec j40_data_pipeline_1 python3 application.py --help"`
-
+- Clean up the census data directories: `docker exec j40_data_pipeline_1 python3 application.py census-cleanup"`
-If you want to visualize logs while running a command, the following temporary workaround can be used:
+- Clean up the data directories: `docker exec j40_data_pipeline_1 python3 application.py data-cleanup"`
-
+- Generate census data: `docker exec j40_data_pipeline_1 python3 application.py census-data-download"`
-- Run `docker-compose up` on the root of the repo
+- Run all ETL processes: `docker exec j40_data_pipeline_1 python3 application.py etl-run"`
-- Open a new tab on your terminal
+- Generate Score: `docker exec j40_data_pipeline_1 python3 application.py score-run"`
-- Then run any command for the application using this format: `docker exec j40_score_1 python3 application.py [command]`
 
 ## Local development
 
-You can run the Python code locally without Docker to develop, using Poetry. However, to generate the census data you will need the [GDAL library](https://github.com/OSGeo/gdal) installed locally. Also to generate tiles for a local map, you will need [Mapbox tippeanoe](https://github.com/mapbox/tippecanoe)
+You can run the Python code locally without Docker to develop, using Poetry. However, to generate the census data you will need the [GDAL library](https://github.com/OSGeo/gdal) installed locally. Also to generate tiles for a local map, you will need [Mapbox tippeanoe](https://github.com/mapbox/tippecanoe). Please refer to the repos for specific instructions for your OS.
 
-Note: If you are using Windows, please follow [these instructions](https://stackoverflow.com/questions/56958421/pip-install-geopandas-on-windows) to install Geopandas locally. If you want to install TippeCanoe, [follow these instrcutions](https://github.com/GISupportICRC/ArcGIS2Mapbox#installing-tippecanoe-on-windows).
+### Windows Users
+- If you want to download Census data or run tile generation, please install TippeCanoe [following these instrcutions](https://github.com/GISupportICRC/ArcGIS2Mapbox#installing-tippecanoe-on-windows).
+- If you want to generate tiles, you need some pre-requisites for Geopandas as specified in the Poetry requirements. Please follow [these instructions](https://stackoverflow.com/questions/56958421/pip-install-geopandas-on-windows) to install the Geopandas dependency locally.
+
+### Setting up Poetry
 
 - Start a terminal
+- Change to this directory (`/data/data-pipeline`)
 - Make sure you have Python 3.9 installed: `python -V` or `python3 -V`
 - We use [Poetry](https://python-poetry.org/) for managing dependencies and building the application. Please follow the instructions on their site to download.
 - Install Poetry requirements with `poetry install`
@@ -125,7 +125,7 @@ Note: If you are using Windows, please follow [these instructions](https://stack
 
 - Make sure you have Docker running in your machine
 - Start a terminal
-- Change to this directory (i.e. `cd score`)
+- Change to this directory (i.e. `cd data/data-pipeline`)
 - If you want to clear out all data and tiles from all directories, you can run: `poetry run python application.py data-cleanup`.
 - Then run `poetry run python application.py census-data-download`
   Note: Census files are not kept in the repository and the download directories are ignored by Git
@@ -137,18 +137,18 @@ Note: If you are using Windows, please follow [these instructions](https://stack
 ### Serve the map locally
 
 - Start a terminal
-- Change to this directory (i.e. `cd score`)
+- Change to this directory (i.e. `cd data/data-pipeline`)
 - Run: `docker run --rm -it -v ${PWD}/data/tiles:/data -p 8080:80 maptiler/tileserver-gl`
 
 ### Running Jupyter notebooks
 
 - Start a terminal
-- Change to this directory (i.e. `cd score`)
+- Change to this directory (i.e. `cd data/data-pipeline`)
 - Run `poetry run jupyter notebook`. Your browser should open with a Jupyter Notebook tab
 
 ### Activating variable-enabled Markdown for Jupyter notebooks
 
-- Change to this directory (i.e. `cd score`)
+- Change to this directory (i.e. `cd data/data-pipeline`)
 - Activate a Poetry Shell (see above)
 - Run `jupyter contrib nbextension install --user`
 - Run `jupyter nbextension enable python-markdown/main`

data/data-pipeline/etl/score/etl_score_post.py

@@ -20,8 +20,8 @@ class PostScoreETL(ExtractTransformLoad):
         self.STATE_CSV = (
             self.DATA_PATH / "census" / "csv" / "fips_states_2010.csv"
         )
-        self.SCORE_CSV = self.SCORE_CSV_PATH / "full" / "usa.csv"
+        self.FULL_SCORE_CSV = self.SCORE_CSV_PATH / "full" / "usa.csv"
-        self.COUNTY_SCORE_CSV = self.SCORE_CSV_PATH / "full" / "usa-county.csv"
+        self.TILR_SCORE_CSV = self.SCORE_CSV_PATH / "tile" / "usa.csv"
 
         self.TILES_SCORE_COLUMNS = [
             "GEOID10",
@@ -59,7 +59,9 @@ class PostScoreETL(ExtractTransformLoad):
         self.states_df = pd.read_csv(
             self.STATE_CSV, dtype={"fips": "string", "state_code": "string"}
         )
-        self.score_df = pd.read_csv(self.SCORE_CSV, dtype={"GEOID10": "string"})
+        self.score_df = pd.read_csv(
+            self.FULL_SCORE_CSV, dtype={"GEOID10": "string"}
+        )
 
     def transform(self) -> None:
         logger.info(f"Transforming data sources for Score + County CSV")
@@ -98,11 +100,9 @@ class PostScoreETL(ExtractTransformLoad):
         del self.score_county_state_merged["GEOID_OTHER"]
 
     def load(self) -> None:
-        logger.info(f"Saving Score + County CSV")
+        logger.info(f"Saving Full Score CSV with County Information")
         self.SCORE_CSV_PATH.mkdir(parents=True, exist_ok=True)
-        # self.score_county_state_merged.to_csv(
+        self.score_county_state_merged.to_csv(self.FULL_SCORE_CSV, index=False)
-        #     self.COUNTY_SCORE_CSV, index=False
-        # )
 
         logger.info(f"Saving Tile Score CSV")
         # TODO: check which are the columns we'll use

data/data-pipeline/etl/sources/census/etl_utils.py

@@ -3,7 +3,14 @@ import csv
 import os
 from config import settings
 
-from utils import remove_files_from_dir, remove_all_dirs_from_dir, unzip_file_from_url
+from utils import (
+    remove_files_from_dir,
+    remove_all_dirs_from_dir,
+    unzip_file_from_url,
+    get_module_logger,
+)
+
+logger = get_module_logger(__name__)
 
 
 def reset_data_directories(data_path: Path) -> None:
@@ -27,6 +34,7 @@ def get_state_fips_codes(data_path: Path) -> list:
 
     # check if file exists
     if not os.path.isfile(fips_csv_path):
+        logger.info(f"Downloading fips from S3 repository")
         unzip_file_from_url(
             settings.AWS_JUSTICE40_DATA_URL + "/Census/fips_states_2010.zip",
             data_path / "tmp",

data/data-pipeline/etl/sources/ejscreen/etl.py

@@ -9,7 +9,7 @@ logger = get_module_logger(__name__)
 class EJScreenETL(ExtractTransformLoad):
     def __init__(self):
         self.EJSCREEN_FTP_URL = "https://gaftp.epa.gov/EJSCREEN/2019/EJSCREEN_2019_StatePctile.csv.zip"
-        self.EJSCREEN_CSV = self.TMP_PATH / "EJSCREEN_2019_StatePctile.csv"
+        self.EJSCREEN_CSV = self.TMP_PATH / "EJSCREEN_2019_StatePctiles.csv"
         self.CSV_PATH = self.DATA_PATH / "dataset" / "ejscreen_2019"
         self.df: pd.DataFrame
 

data/data-pipeline/etl/sources/housing_and_transportation/etl.py

@@ -25,12 +25,19 @@ class HousingTransportationETL(ExtractTransformLoad):
             logger.info(
                 f"Downloading housing data for state/territory with FIPS code {fips}"
             )
+
+            # Puerto Rico has no data, so skip
+            if fips == "72":
+                continue
+
             unzip_file_from_url(
                 f"{self.HOUSING_FTP_URL}{fips}", self.TMP_PATH, zip_file_dir
             )
 
             # New file name:
-            tmp_csv_file_path = zip_file_dir / f"htaindex_data_blkgrps_{fips}.csv"
+            tmp_csv_file_path = (
+                zip_file_dir / f"htaindex_data_blkgrps_{fips}.csv"
+            )
             tmp_df = pd.read_csv(filepath_or_buffer=tmp_csv_file_path)
 
             dfs.append(tmp_df)
@@ -44,9 +51,9 @@ class HousingTransportationETL(ExtractTransformLoad):
 
         # Rename and reformat block group ID
         self.df.rename(columns={"blkgrp": self.GEOID_FIELD_NAME}, inplace=True)
-        self.df[self.GEOID_FIELD_NAME] = self.df[self.GEOID_FIELD_NAME].str.replace(
+        self.df[self.GEOID_FIELD_NAME] = self.df[
-            '"', ""
+            self.GEOID_FIELD_NAME
-        )
+        ].str.replace('"', "")
 
     def load(self) -> None:
         logger.info(f"Saving Housing and Transportation Data")

data/data-pipeline/etl/sources/hud_housing/etl.py

@@ -11,16 +11,16 @@ class HudHousingETL(ExtractTransformLoad):
     def __init__(self):
         self.OUTPUT_PATH = self.DATA_PATH / "dataset" / "hud_housing"
         self.GEOID_TRACT_FIELD_NAME = "GEOID10_TRACT"
-        self.HOUSING_FTP_URL = (
+        self.HOUSING_FTP_URL = "https://www.huduser.gov/portal/datasets/cp/2012thru2016-140-csv.zip"
-            "https://www.huduser.gov/portal/datasets/cp/2012thru2016-140-csv.zip"
-        )
         self.HOUSING_ZIP_FILE_DIR = self.TMP_PATH / "hud_housing"
 
         # We measure households earning less than 80% of HUD Area Median Family Income by county
         # and paying greater than 30% of their income to housing costs.
         self.HOUSING_BURDEN_FIELD_NAME = "Housing burden (percent)"
         self.HOUSING_BURDEN_NUMERATOR_FIELD_NAME = "HOUSING_BURDEN_NUMERATOR"
-        self.HOUSING_BURDEN_DENOMINATOR_FIELD_NAME = "HOUSING_BURDEN_DENOMINATOR"
+        self.HOUSING_BURDEN_DENOMINATOR_FIELD_NAME = (
+            "HOUSING_BURDEN_DENOMINATOR"
+        )
 
         # Note: some variable definitions.
         # HUD-adjusted median family income (HAMFI).
@@ -47,10 +47,15 @@ class HudHousingETL(ExtractTransformLoad):
             / "140"
             / "Table8.csv"
         )
-        self.df = pd.read_csv(filepath_or_buffer=tmp_csv_file_path)
+        self.df = pd.read_csv(
+            filepath_or_buffer=tmp_csv_file_path,
+            encoding="latin-1",
+        )
 
         # Rename and reformat block group ID
-        self.df.rename(columns={"geoid": self.GEOID_TRACT_FIELD_NAME}, inplace=True)
+        self.df.rename(
+            columns={"geoid": self.GEOID_TRACT_FIELD_NAME}, inplace=True
+        )
 
         # The CHAS data has census tract ids such as `14000US01001020100`
         # Whereas the rest of our data uses, for the same tract, `01001020100`.
@@ -160,7 +165,9 @@ class HudHousingETL(ExtractTransformLoad):
         # TODO: add small sample size checks
         self.df[self.HOUSING_BURDEN_FIELD_NAME] = self.df[
             self.HOUSING_BURDEN_NUMERATOR_FIELD_NAME
-        ].astype(float) / self.df[self.HOUSING_BURDEN_DENOMINATOR_FIELD_NAME].astype(
+        ].astype(float) / self.df[
+            self.HOUSING_BURDEN_DENOMINATOR_FIELD_NAME
+        ].astype(
             float
         )
 

data/data-pipeline/pyproject.toml

@@ -6,16 +6,16 @@ authors = ["Your Name <you@example.com>"]
 
 [tool.poetry.dependencies]
 python = "^3.7.1"
+CensusData = "^1.13"
+click = "^8.0.1"
+dynaconf = "^3.1.4"
 ipython = "^7.24.1"
 jupyter = "^1.0.0"
 jupyter-contrib-nbextensions = "^0.5.1"
 numpy = "^1.21.0"
 pandas = "^1.2.5"
 requests = "^2.25.1"
-click = "^8.0.1"
-dynaconf = "^3.1.4"
 types-requests = "^2.25.0"
-CensusData = "^1.13"
 
 [tool.poetry.dev-dependencies]
 mypy = "^0.910"

docker-compose.yml

@@ -1,13 +1,13 @@
 version: "3.4"
 services:
   score:
-    image: j40_score
+    image: j40_data_pipeline
-    container_name: j40_score_1
+    container_name: j40_data_pipeline_1
-    build: score
+    build: data/data-pipeline
     ports:
       - 8888:8888
     volumes:
-      - ./score:/score
+      - ./data/data-pipeline:/data-pipeline
     stdin_open: true
     tty: true
     environment: