Problem Statement
During El Niño episodes in the Philippines, agricultural damage routinely exceeds ₱6 to ₱10 billion, devastating over 150,000 farmers across Western Visayas, Cagayan Valley, Central Luzon, and Northern Mindanao. The National Irrigation Administration (NIA) is forced to shut off irrigation canals to prioritize municipal drinking water, leaving rice fields cracked and parched. Smallholder farmers—unaware of the severity or timeline of impending drought—continue planting water-intensive traditional paddy rice, losing their entire investment in seeds, fertilizer, and fuel for irrigation pumps. Farmers lack access to actionable agrometeorological guidance that tells them what alternative crops can survive on residual soil moisture while fetching viable market prices.
Proposed Solution
Tanim-Init is an AI agronomist and climate-adaptation advisor built specifically for Filipino farmers facing El Niño droughts.
- Drought Risk & Crop-Switching Engine: Ingesting multi-spectral satellite imagery (Sentinel-2 NDVI, NASA SMAP soil moisture) and PAGASA climate forecasts, the AI calculates a 90-day field-level drought severity index. If irrigation failure is projected, it generates a personalized crop-switch recommendation—advising farmers to transition from paddy rice to drought-resistant crops like mungbean (balatong), sorghum, peanuts, or sweet potato, complete with planting window calendars and projected farm-gate profit margins.
- Precision Water Allocation: For farmers with shallow tube wells or rainwater storage, the AI calculates evapotranspiration rates to provide an exact drip/furrow irrigation schedule, conserving up to 45% of groundwater.
- Dialect-First Low-Bandwidth Delivery: Farmers do not need smartphones; weekly guidance and emergency pest/heat alerts are sent via automated Smart SMS or voice calls in Ilocano, Tagalog, Hiligaynon, or Cebuano.
- LGU & DA Dashboard: Municipal Agriculturists and the Department of Agriculture (DA) track a live heatmap of drought-stressed acreage and crop diversification adoption across towns.
Target Market
- Primary Users: ~2.5 million Filipino smallholder rice, corn, and vegetable farmers vulnerable to El Niño droughts.
- Buyers & Partners (B2G / B2B): Department of Agriculture (DA), National Irrigation Administration (NIA), PhilRice, Provincial/Municipal Agriculture Offices, agricultural cooperatives, and microfinance institutions insuring farmers.
How AI Is Used
- Satellite Soil Moisture & Crop Health Analysis: Computer vision and spectral indexing (NDVI, NDWI, EVI) monitor real-time crop water stress and vegetation degradation across agrarian reform communities.
- Predictive Agronomic Recommendation: Machine learning models trained on Philippine agro-climatic zones, soil types, and Bureau of Plant Industry (BPI) crop databases match projected moisture deficits to optimal alternative crop varieties.
- Conversational Dialect Interface: Speech-to-speech and multilingual LLMs translate complex agrometeorological data into practical, respectful instructions in regional dialects.
- Why it fails without AI: General weather forecasts lack field-level granularity; without satellite vision and predictive soil modeling, extension workers cannot manually inspect millions of dispersed family farms in time to prevent planting failures.
PLDT / Smart / ePLDT / Rokid Angle
- Smart Communications (Rural Connectivity & SMS): Smart’s extensive rural cellular footprint and SMS infrastructure deliver alerts directly to keypad phones in remote farming communities.
- ePLDT VITRO Cloud: Securely stores national geospatial farm parcel maps and trains machine learning models on high-volume satellite imagery.
- PLDT Enterprise Sustainability (ESG): Positioned as PLDT Enterprise’s flagship digital agriculture and food security sustainability initiative, aligning with Philippine national climate action goals.
Demo Scenario (3 minutes for judges)
- The Stranded Farmer (0:00 - 0:45): A split screen introduces “Mang Lando”, a rice farmer in Nueva Ecija. NIA announces that the Pantabangan canal will cut off water in 2 weeks due to El Niño. Without intervention, his ₱40,000 rice seed investment will wither.
- The Satellite & AI Diagnosis (0:45 - 1:30): The presenter enters Mang Lando’s barangay into Tanim-Init. The AI displays Sentinel-2 moisture depletion maps, flags an 85% probability of drought catastrophe for paddy rice, and computes the optimal switch: Yellow Mungbean (Pag-asa 7 variety), which matures in 60 days on minimal residual moisture.
- The Voice/SMS Alert (1:30 - 2:15): The demo plays an audio call simulation in Tagalog/Ilocano: “Mang Lando, dahil sa El Niño, huwag munang magpunla ng palay. Magtanim ng munggo bago matapos ang Oktubre. May residual moisture pa ang lupa, at may guaranteed buyer ang cooperative sa halagang ₱65/kilo.”
- DA Municipal Dashboard (2:15 - 3:00): The presenter displays the Municipal Agriculture Officer’s dashboard showing 450 farmers adopting the crop-switch protocol, preventing ₱18 million in agricultural losses and conserving 30 million liters of groundwater.
30-Hour Build Plan
| Hours | Task | Role |
|---|---|---|
| 0–4 | Git repo, ingest Sentinel-2/Landsat sample GeoTIFFs for Central Luzon, download PhilRice crop water requirement guidelines | All |
| 4–12 | Build the drought severity scoring model and crop-recommendation engine in Python (Rasterio, GeoPandas, Scikit-Learn) | AI / Data Lead |
| 12–20 | Build the Municipal Agriculture dashboard with interactive farm stress maps (React + Leaflet) | Frontend Lead |
| 12–20 | Develop the multilingual SMS/voice generation engine translating crop advice into Tagalog and Ilocano | Backend Lead |
| 20–26 | End-to-end simulation of a 60-day El Niño drought scenario across 3 farming municipalities; benchmark water savings | Full Team |
| 26–30 | Pitch deck, farmer testimonial video snippet, rehearse 3-minute pitch | Pitcher & UI |
Suggested stack: Python (FastAPI, Rasterio, GeoPandas, Xarray), React / Next.js, Tailwind CSS, Leaflet/Mapbox, Twilio/Semaphore API (SMS), ElevenLabs or open-source TTS for Filipino voice, SQLite/PostgreSQL.
Public data / APIs: Copernicus Open Access Hub (Sentinel-2 NDVI), NASA POWER Agroclimatology API, PAGASA El Niño Bulletins, PhilRice Rice Science Portal open datasets.
Risks & Privacy Notes
- Agronomic Liability: Crop switching carries market and disease risks. Recommendations must include clear disclaimers that they are advisories developed with DA-BPI protocols, not financial guarantees.
- Privacy (Clause 24 Compliance): Only farmer phone numbers and approximate farm cluster GPS coordinates are processed. No land titles, banking credentials, or personal identification cards are stored.
AI Use Disclosure (draft)
Tanim-Init uses satellite spectral index calculation (NDVI/NDWI) combined with random forest classifiers to evaluate crop water stress. LLMs and neural TTS tools assist in generating localized, dialect-appropriate SMS and audio transcripts. The system incorporates agronomic guidelines published by PhilRice and the Department of Agriculture.
Rubric Scoring
| Criterion | Weight | Score (1–5) | Why |
|---|---|---|---|
| Innovation & creativity | 25% | 4.4 | Shifts disaster response from post-calamity financial relief (ayuda) to proactive, predictive crop adaptation. |
| Problem–solution fit | 25% | 4.7 | Tackles the direct cause of billions in agricultural damage and rural food insecurity during recurring dry spells. |
| AI application & value | 20% | 4.5 | Combines multi-spectral satellite remote sensing with localized agronomic expert modeling that cannot be done manually at scale. |
| Feasibility | 15% | 4.2 | Leverages freely available Sentinel/NASA satellite data and simple SMS channels accessible to every Filipino farmer. |
| Impact & scalability | 15% | 4.6 | Highly scalable across millions of agricultural workers in the Philippines and Southeast Asia; strong government partnership appeal. |
| Weighted total | 4.50 | (0.25*4.4) + (0.25*4.7) + (0.20*4.5) + (0.15*4.2) + (0.15*4.6) = 4.50 |