Apppublic

harri620/OhioFoodInsecurity

sourceHugging Facemitupdated 4mo agoView on Hugging Face

0likes

App README

Food Insecurity Data Pipeline

Project Overview

This project analyzes food insecurity trends across U.S. counties and states, with a focused comparison of Ohio against the national average. It combines three real-world data sources to build a single tidy, validated dataset ready for Tableau or Power BI visualization.

Research Question: How does food insecurity in Ohio compare to the rest of the United States, and do county-level patterns within Ohio mirror or diverge from broader U.S. trends?

Data Sources

#	Source	Acquisition Method	Coverage	Key Variables
1	`MMG2025_2019-2023_Data_To_Share.xlsx`	Excel file load	County & State, 2019–2023	Food insecurity rate, # food insecure, child FI rate, meal cost
2	`frac.org/hunger-poverty-america`	Web scraping (`rvest`)	National summary	National FI rate, state FI range, poverty/SNAP context
3	U.S. Census Bureau ACS 5-Year	API (`tidycensus`)	County & State, 2023	Poverty rate, median income, unemployment rate, population

Note on scraping target: feedingamerica.org/hunger-in-america/[state] pages are JavaScript-rendered — rvest reads raw HTML before JS executes, so all stat values return blank. FRAC's hunger-poverty-america page is static HTML and fully scrapeable with rvest.

Repository Structure

food_insecurity_pipeline/
│
├── README.md                          # This file
├── .gitignore                          # Excludes credentials, data files, R artifacts
│
├── 01_data_acquisition/
│   ├── 01a_load_mmg_excel.R            # Loads MMG2025 Excel (county + state sheets)
│   ├── 01b_scrape_frac.R               # Scrapes FRAC hunger-poverty-america page
│   └── 01c_census_api.R                # Pulls ACS data via tidycensus API
│
├── 02_cleaning/
│   ├── 02a_clean_mmg.R                 # Cleans MMG: FIPS padding, type coercion, flags
│   ├── 02b_clean_scraped.R             # Cleans FRAC scraped data, parses percentages
│   └── 02c_clean_census.R              # Cleans Census: derived poverty/unemployment rates
│
├── 03_merge/
│   └── 03_merge_final.R                # Merges all three sources on FIPS + state name
│
├── 04_validation/
│   └── 04_validate.R                   # 15 data quality checks; outputs validation_table.csv
│
├── 05_eda/
│   └── 05_eda_summaries.R              # Summary stats, Ohio vs. U.S. tables, 4 ggplot visuals
│
└── data/
    ├── raw/                            # Raw and intermediate .rds files (not committed)
    └── final/                          # Final merged dataset + validation table + EDA outputs

How to Reproduce the Pipeline

Prerequisites

Install required R packages:

install.packages(c(
  "tidyverse", "readxl", "rvest", "stringr",
  "tidycensus", "janitor", "here", "skimr",
  "ggplot2", "scales", "writexl"
))

Set Up the Census API Key Securely

Never hard-code your API key in any script.

1.Get a free key at: https://api.census.gov/data/key_signup.html
2.Open your .Renviron file:

usethis::edit_r_environ()

1.Add this line and save:

CENSUS_API_KEY=your_actual_key_here

1.Restart R. Your key is now available via Sys.getenv("CENSUS_API_KEY").

⚠️ .Renviron is listed in .gitignore and will never be committed to GitHub.

Place the MMG Excel File

Download MMG2025_2019-2023_Data_To_Share.xlsx from Feeding America and place it at:

data/raw/MMG2025_2019-2023_Data_To_Share.xlsx

Run the Pipeline (in order)

source("01_data_acquisition/01a_load_mmg_excel.R")
source("01_data_acquisition/01b_scrape_frac.R")
source("01_data_acquisition/01c_census_api.R")
source("02_cleaning/02a_clean_mmg.R")
source("02_cleaning/02b_clean_scraped.R")
source("02_cleaning/02c_clean_census.R")
source("03_merge/03_merge_final.R")
source("04_validation/04_validate.R")
source("05_eda/05_eda_summaries.R")

Final Dataset Structure

File: data/final/food_insecurity_final.csv

Column	Source	Description
`fips`	MMG	5-digit county FIPS code
`county_name`	MMG	County name
`state_name`	MMG	State name
`year`	MMG	Data year (2019–2023)
`is_ohio`	Derived	TRUE if Ohio county
`fi_rate`	MMG	Overall food insecurity rate
`fi_count`	MMG	Number of food insecure persons
`child_fi_rate`	MMG	Child food insecurity rate
`meal_cost`	MMG	Average meal cost ($)
`poverty_rate`	Census ACS	Poverty rate (derived)
`unemployment_rate`	Census ACS	Unemployment rate (derived)
`median_income`	Census ACS	Median household income
`total_population`	Census ACS	Total county population
`frac_national_fi_pct`	FRAC scrape	National FI rate reference
`frac_state_fi_min_pct`	FRAC scrape	Lowest state FI rate
`frac_state_fi_max_pct`	FRAC scrape	Highest state FI rate
`_flag / _invalid`	Derived	Data quality flags

Data Quality Notes

—Missing values: Flagged with missing_* boolean columns; documented in validation_table.csv
—FIPS standardization: All FIPS codes zero-padded to 5 digits for consistent joining
—Duplicate handling: Deduplicated on fips + year (MMG) and state_name (FRAC) — first valid row kept
—Census join coverage: County Census data joined for 2023 ACS and applied as context across all MMG years
—Impossible values: Rates outside 0–100%, negative counts, and zero/negative incomes are flagged in the validation table

Security

—API keys are stored in .Renviron (excluded via .gitignore)
—No credentials, tokens, or secrets appear anywhere in any script
—Raw Excel and CSV data files are excluded from the repository via .gitignore
—The .gitignore also excludes .RData, .Rhistory, and .Rproj.user/