Share2Care

Connecting donors with distributors

Comment

Inspiration

"Resources are not free; whenever you waste them, someone else somewhere else is paying for it" -Anonymous

"Do not waste water, even if you perform your ablution on the bank of an abundantly-flowing river." — Prophet Muhammad ﷺ (Sunan Ibn Majah, Hadith 425)

What it does

Ensures efficient distribution of edible donations. Prevents wastage of food. Impact: Targets SDG 2 directly, uses live humanitarian data to guide donations where they matter most. Innovation: Combines real-time need mapping with behavioral psychology to turn donors into regular supporters. Feasibility & Scalability: Works immediately for Pakistan floods; easily replicable for other crises with different OCHA datasets.

🛠️ How We Built It

🚀 Overview

Share2Care – Zero Hunger is a Streamlit-based web application that connects surplus food donors, NGOs, and recipients through real-time mapping, AI-driven insights, and behavioral nudges that sustain long-term generosity.

🗺️ Core Features

1. Interactive Food Insecurity Map

  • Visualizes OCHA 5W and Pakistan Administrative Boundaries data.
  • Integrates IPC acute food insecurity overlays.
  • Built using Folium + GeoPandas for spatial joins and choropleth maps.

2. Donor ↔ NGO Matching

  • Simple pledge and claim form flow stored locally via CSV/SQLite for MVP.
  • NGOs and volunteers can claim nearby pledges in real-time.

3. Food Price Monitoring & Forecasting

  • Uses WFP food price dataset for commodity trends.
  • Forecasts shocks using Prophet or fallback ARIMA.
  • Generates early affordability alerts.

4. Psychology Layer (Nudges & Reflection)

  • Sentiment analysis on donor notes using DistilBERT.
  • Provides empathetic feedback, micro-commitments, and positive reinforcement.
  • Encourages repeat donations through small rewards and reminders.

5. Food Image Tagging (Optional but integrated)

  • Uses MobileNetV2 to auto-classify uploaded food images.
  • Filters items by curated food keywords for better matching.

🤖 AI Components

Component Model / Library Purpose
Sentiment Analysis DistilBERT / TextBlob Analyze donor mood & guide nudges
Food Image Tagging MobileNetV2 + OpenCV Classify uploaded food images
Price Forecasting Prophet / ARIMA Predict food price surges
Mood Logging NLP Sentiment + Mood Logs Track donor engagement & emotion trends
Motivational Nudges Custom Psychology Module Reinforce recurring altruistic behavior

🧩 Technical Architecture

Backend:

  • Built with FastAPI and Python.

  • Folder structure:

    • config.py – defines data paths.
    • data_loader.py – loads and merges OCHA, IPC, and WFP data.
    • models/ – includes image_tagging.py, price_forecast.py, sentiment.py.

Frontend:

  • React app for an interactive interface.
  • Local data persistence via CSV / SQLite for MVP simplicity.

💡 Data Sources

Dataset Purpose Source
OCHA Humanitarian Response (5W) Severity by admin area; NGO presence humdata.org
Pakistan Admin Boundaries Choropleth mapping, tooltips GADM / OCHA GeoJSONs
IPC Food Insecurity Data Acute food insecurity overlays IPC Info
WFP Food Price Monitoring Historical + forecast data data.humdata.org
NGO Directories (ReliefWeb / HDX) Matching donors with local NGOs reliefweb.int/organizations

🧠 Psychology & Behavior Layer

  • Altruism Boost: Public acknowledgment of contributions (e.g., “You helped 3 families today”).
  • Habit Loop: Micro-commitments like “Pledge one meal this week” and reminders.
  • Behavioral Nudges: Local short-term goals (e.g., “10 meals needed today in Muzaffargarh”).
  • Mood Reflection: Sentiment-based supportive messages (positive reinforcement / empathy).

⚡ MVP Scope

  • Interactive Folium map for food insecurity visualization.
  • Minimal donor–NGO form flow.
  • Price forecasting with Prophet/ARIMA fallback.
  • Sentiment-driven nudges & mood tracking.
  • Optional image tagging using MobileNetV2.
  • Local data storage & CSV exports for hackathon submission.

🎯 Hackathon Pitch Summary

  • Impact: Directly contributes to SDG 2 – Zero Hunger.
  • Innovation: Combines AI, real-time data, and behavioral psychology.
  • Feasibility: Fully functional MVP for Pakistan floods; adaptable to other crises.
  • Scalability: Extendable to droughts, earthquakes, or refugee contexts globally.

Challenges we ran into

  • One of our partners got sick so the workload increased for the rest of us
  • Having a small team made it slightly difficult for us to complete it in time

Accomplishments that we're proud of

  • Completing it for submission just-in-time was itself an accomplishment for us
  • We're proud that we developed something which has a practical implementation and the potential to help at least thousands, if not millions, of people living in adversity and food scarcity.

What we learned

  • Using publicly available data for the good of the people
  • The actual potential of what we just developed
  • Different ways data and tools can be tweaked to help the people in need, even in areas you cannot reach yourself and the people you do not know.

What's next for Share2Care

  • Hopefully, a national integration and acceptance by the main NGOs, Government, Restaurants, Caterers, and other entities which can act as Donors.
  • Practical implementation of this to really help out the people in need, not just the flood-affected people, but also the people living in food-scarce areas with malnourished food supplies.
  • Mobile App to enhance user experiece and Accessibility
  • Food categories according to shelf life
  • Two side confirmation of donation delivery reducing the admin role
  • Communities or Peoples records for receiving data ensuring uniform distributions accordingg to community needs
  • Admin approval over releasing donor information to reduce spamming and enhance data privacy and protection *Heat maps and high demand indications for location and resources ensuring optimal use of donations/resources.
  • Module to collect spoiled food and recycle or use it for fertilizers.
  • Developing it as a Non Profit platform, and extending it beyond food to other necessities. *Expanding it to an online wall of kindness
  • Launching awareness camapaigns

Built With

Share this project:

Updates