This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
CHECK OUR OPEN APP HERE: OpenTrips
Team: HONEYBEE
- Ashwin Temkar (@ashwintemkar ) and Kiranraj Chintada (@kiranrajchintada302 ).
What We Built
OpenTrips answers one question before you leave the house. Should I go out this weekend?
You type a place, two dates, and a sentence about yourself. "Beginner, bad knee, budget 3000 rupees, wants hills." You get three things back.
- A go, caution, or wait verdict for rain and landslide risk, with the model's measured accuracy on the screen.
- A day by day outdoor plan and a packing list from a model we fine tuned for that job.
- A map of recorded landslides nearby, the forecast grid, and places to stay, so you can leave with the plan instead of researching it.
The screen is the planning step. The point of the plan is to close the laptop. It is for anyone planning a walk, a day in the hills, or a camp, anywhere a forecast exists. It is not a hotel trip planner. The itinerary is trails, viewpoints, and short walks sized to ability. Stays on the map are context, not the product.
| Product | Outdoor go or wait, plus a plan and a packing list |
| Risk model | TabPFN on real NASA landslide events |
| Plan model | Qwen3 8B, fine tuned on Tinker |
| Weather and rain history | Open Meteo, no API key |
| Maps | OpenStreetMap and Leaflet |
| Hosting | Render, API and web app |
| Saved plans | Postgres, reopen by link, cached for no signal |
| Split | Majority guess | Logistic regression | TabPFN |
|---|---|---|---|
| Time, last 20 percent of each class | 0.556 | 0.799 | 0.914 accuracy, 0.972 AUC |
| Geography, India-tagged rows held out | 0.390 | 0.821 | 0.899 accuracy, 0.956 AUC |
The plan model, on 10 requests it never saw in training: base model 0 of 10, fine tuned model 8 of 10.
The 0.390 majority guess always says non-event, because that is the most common training label outside India. Always saying event on the India rows scores 0.610. TabPFN at 0.899 is above both. The India rows in the csv are 775 events and 496 non-events. The plan result is a small sample. A 95 percent Wilson interval on 8 of 10 is about 49 percent to 94 percent.
flowchart TD
A[Place, dates, one sentence] --> B[OpenStreetMap coordinates]
B --> C[TabPFN risk]
B --> D[Qwen3 8B plan]
C --> E[Go, caution, or wait plus accuracy]
D --> F[Days, packing, warnings]
F --> G[Rule checker]
G --> H[Saved link]
H --> I[Leave, plan cached for no signal]
Demo
Video: ADDING_SOON_THE_YOUTUBE_URL
Live app: OpenTrips
The API on a free Render instance can take about half a minute to wake up on the first request.
Screenshots
What the demo should show, in order:
1. Type a place, dates, and a constraint sentence.
2. The verdict appears with accuracy 0.899, not a buried footnote.
3. The day list, packing list, and whether the constraint check passed.
4. The map. Recorded landslides, rain grid, stays.
5. Refresh the saved link. The plan is still there.
Code
ashwintemkar
/
opentrips
Should I go out this weekend? Rain/landslide risk + AI trip planning, anywhere in the world. Built for Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass.
OpenTrips
Should I go out this weekend?
You name a place, pick dates, and describe yourself in one sentence. OpenTrips returns a rain and landslide risk verdict with the model's measured accuracy, plus a day by day outdoor plan and a packing list. The plan is for walks, hills, and camps. It is not a hotel itinerary.
Built for the Hacktoberfest Open Source AI Challenge, Week 1, Touch Grass.
| Code | https://github.com/ashwintemkar/opentrips |
| Risk model | TabPFN on real NASA landslide events |
| Plan model | Qwen3 8B, fine tuned with Tinker |
| Weather | Open Meteo, no API key |
| Maps | OpenStreetMap |
What you get
- A verdict of go, caution, or wait, and the accuracy number 0.899 on the screen.
- Days, distances, a packing list, and a pass or fail check against your budget and ability.
- A map of recorded landslides nearby, the forecast grid, and places to stay.
- A link you can reopen. A plan you…
https://github.com/ashwintemkar/opentrips
| File | What a judge can check |
|---|---|
landslide_dataset.csv |
3,935 training rows, events and non events, rainfall features |
train_tabpfn.py |
Time split and India held out split, versus majority and logistic regression |
add_india_negatives.py |
Random India non events, so the India test has both classes |
train_plan_model.py |
Tinker fine tune of Qwen3 8B |
compare_plans.py |
Base model versus fine tuned model on held out requests |
plan_examples.jsonl |
The 24 outdoor prompt and JSON pairs |
rule_checker.py |
Budget and distance checks, used live by the API |
main.py |
POST /risk and POST /plan
|
frontend/ |
Form, results, map, saved plan page |
How I Built It
Two services. Kiran built the Next.js app people touch. I built the data, TabPFN, the Tinker fine tune, and the FastAPI service. Open Meteo and OpenStreetMap sit in the middle. Neither needs a paid key.
flowchart LR
U[Person] --> WEB[Next.js on Render]
WEB --> GEO[OpenStreetMap place search]
WEB --> RISK[POST /risk]
WEB --> PLAN[POST /plan]
WEB --> MAP[Map and nearby layers]
RISK --> METEO[Open Meteo forecast]
RISK --> TAB[TabPFN on NASA landslide rows]
PLAN --> QWEN[Qwen3 8B fine tuned on Tinker]
MAP --> NASA[NASA events near the pin]
WEB --> DB[Postgres saved plan link]
sequenceDiagram
participant P as Person
participant W as Web app
participant R as Risk API
participant L as Plan model
P->>W: Place, dates, constraints
W->>W: Resolve the place on OpenStreetMap
W->>R: Latitude, longitude, dates
R->>R: Rain windows from Open Meteo
R-->>W: Risk, verdict, accuracy 0.899
W->>L: The sentence plus that risk
L-->>W: Days, packing, warnings, checks
W-->>P: Plan link they can reopen
TabPFN, on recorded events
The training table is the NASA Global Landslide Catalog plus rainfall from the Open Meteo archive. The raw export has 11,033 events in 142 countries. We kept events located within 10 km. 786 India events passed that filter. 775 of those have rainfall and are in the csv. We sampled the rest of the world down for runtime. Each event gets 1 day, 3 day, 7 day, and 14 day rainfall ending on the event date. Non events are random places and dates run through the same rainfall pipeline. The 496 India non events come from add_india_negatives.py, so the India test has both classes. They are not rainy days at known landslide places with no recorded event. A cut on 14 day rain alone scores 0.819 on India, close to the logistic regression.
landslide_dataset.csv has 3,935 rows. It is committed so the result is not a claim judges have to take on faith.
flowchart TD
A[NASA catalog, 11033 events, 142 countries] --> B[Keep location accuracy within 10 km]
B --> C[775 India events with rain in the csv]
B --> D[960 events sampled from the rest of the world]
C --> E[Open Meteo archive rain]
D --> E
E --> F[1, 3, 7, and 14 day rainfall]
G[Random non events, same rain features] --> H[Plus 496 India non events]
F --> I[landslide_dataset.csv, 3935 rows]
H --> I
I --> J[TabPFN fit at API startup]
flowchart LR
subgraph timeSplit [Time split]
T1[Older rows of each class] --> T2[Train]
T3[Last 20 percent of each class] --> T4[Test]
end
subgraph geoSplit [India held out]
G1[Every country except India] --> G2[Train]
G3[India events and India non events only] --> G4[Test]
end
T2 --> M[TabPFN versus majority versus logistic regression]
T4 --> M
G2 --> M
G4 --> M
flowchart TD
R[Live forecast rain plus latitude and longitude] --> P[TabPFN probability]
P --> W{Score}
W -->|above 0.6| WAIT[wait]
W -->|above 0.3| CAUTION[caution]
W -->|otherwise| GO[go]
WAIT --> UI[Shown with accuracy 0.899]
CAUTION --> UI
GO --> UI
At request time the API pulls the live forecast for those same rainfall windows and asks TabPFN for the probability of a landslide style event. Above 0.6 is wait. Above 0.3 is caution. Otherwise go. The number on the card is the geography split, 0.899. The training table is about 44 percent events, far more common than a real weekend, so the 0.6 and 0.3 cuts are not the chance a wet weekend is safe. Precision and recall stay blank, because we did not compute them. We did not make a calibration plot.
The first evaluation runs looked strong and were wrong. One split had no India non events, so the model was mostly saying "event." We fixed the data and published the number from the fixed split. The holdout is the country tag, not a distance buffer. 130 of the 1,271 India test rows are within 10 km of a training event. The median distance is 99 km. The time split is stratified too. The last 20 percent of each class by date, not one global sort that dumped almost all of one class into the test slice.
Tinker, on an open weight model
The plan is not a second call to a general chat model. The base model is Qwen/Qwen3-8B, which is open weight. We fine tuned a LoRA of rank 16 on Tinker.
Training data is 24 hand written prompt and JSON pairs. Places include India, the US, the UK, Switzerland, Nepal, Japan, Peru, New Zealand, Kenya, France, Iceland, and South Africa. Constraints include a bad knee, asthma, an elderly parent, a toddler, fear of heights, and pregnancy. The instruction prefix is the same at training time and at request time.
The first run was a single gradient step. The model still wrote chain of thought and ignored the JSON. A 20 epoch loop fixed that. The model then repeated the answer until the token limit, because it never learned to stop. We keep the first balanced JSON object instead of training again.
A rule checker scores each plan. Valid JSON, at least one day, a packing list, budget at or under the number in the sentence, and distance per day inside a cap inferred from words like beginner or advanced. Held out result:
| Pass rate | |
|---|---|
| Base Qwen3 8B | 0 of 10 |
| Fine tuned checkpoint | 8 of 10, small sample |
The live /plan route runs that checker. checks_passed is the checker result. It is not hardcoded to true.
Checkpoint: tinker://4deeaa51-09a3-5d86-8281-29ad275d2c0c:train:0/sampler_weights/opentrips-plan-v1
What the app does around those two models
The form resolves the place with OpenStreetMap Nominatim, then calls /risk and /plan. The results page shows the verdict, the accuracy, the itinerary, and a Leaflet map. Map layers are nearby NASA landslide points, an Open Meteo weather grid, and stays ranked by distance and by how close they are to a recorded landslide. A plan can be saved and reopened by link. A service worker caches a plan page you already opened, so that page still loads with no signal.
That is the offline behavior. The models do not run on the phone in the forest. The plan is made before you leave, then cached.
Why Does Open Innovation Matter?
A closed chat API can write a pleasant itinerary. It cannot show a measured error rate on landslide records it was never trained on, and you cannot ship the training table next to the code.
| Closed default | What open made possible |
|---|---|
| Ask a hosted chat model if the weekend looks fine | TabPFN on a public CSV of real events |
| Labels the author wrote, then a model that copies them | Labels are recorded reports, with a majority baseline and a logistic regression baseline |
| One country, no held out region | India-tagged rows held out, tested only on those rows, 0.899 accuracy |
| A base model that chats | Fine tuned open weight Qwen3 8B that emits JSON, 8 of 10 versus 0 of 10, small sample |
| Weather behind a paid key | Open Meteo, no key |
| Maps behind a paid vendor | OpenStreetMap |
| Model swap means a new product | The CSV, the checker, and the API stay. The checkpoint can change |
Qwen3 8B weights are public. Tinker let us change them for this task and keep the checkpoint. TabPFN let us treat a few thousand rows as a classifier without standing up a training cluster. The dataset is in the repo so someone else can rerun the comparison.
We do not claim the phone runs the model with no signal. A new plan needs the server, because the forecast and the models run there. The honest offline piece is the saved plan you already opened.
Prize Categories
- Best Use of TabPFN. The risk score is TabPFN on historical landslide rows. Baselines are in the table above. Accuracy 0.899 is on every result.
- Best Use of Tinker. Qwen3 8B fine tuned to write the plan. Held out pass rate 8 of 10 against 0 of 10 for the base model, a small sample.
- Best Use of Render. The FastAPI model service and the Next.js app are both hosted on Render.
Thanks for reading. If the verdict says go, close the tab.





Top comments (3)
Publishing the number from the fixed split and saying what went wrong in the first runs is the right habit. Two things I'd check on the 0.899.
The India majority baseline of 0.390 depends on which class it guesses. If the India test is roughly 786 events and 496 non-events, as the dataset section reads, an all-"event" guess scores about 0.61, and that is the bar 0.899 has to clear. The same split, a different baseline.
The second is that the non-events are random places and dates. Those are mostly dry days, so the model may be learning "was it rainy" rather than "is this slope likely to fail". A better non-event set is rainy days at known landslide-prone spots with no recorded event. Training is also balanced, so the 0.6 and 0.3 cutoffs apply to a prior far above the real one. On a real wet weekend, "wait" may fire a lot more often than the card's accuracy suggests. A calibration plot, or Precision at a fixed alert rate, would answer that better than accuracy.
On the plan model, 8 of 10 has a 95% Wilson interval of roughly 49% to 94%, so I'd report it as "8/10, small sample" and grow the held-out set. Did the 0.899 holdout include locations within a few km of training events?
Thank you for checking the numbers this carefully. You are right on each of these.
The 0.390 baseline does depend on which class it guesses. An all event guess on the India test is about 0.61, and so I consider that is a fair bar to put next to 0.899.
The non-events are random places and dates, so the model can score well by learning that it rained, rather than which slope is likely to fail. A better set would be rainy days at known landslide places with no recorded event. We did not build that set for this challenge.
The training table is much more balanced than a real weekend. The 0.6 and 0.3 cuts were chosen on this table. We have no calibration plot, and no precision at a fixed alert rate. On a wet weekend, wait can show up more often than 0.899 suggests.
On the plan model, the range you give for 8 of 10 is right. "8/10, small sample" is the fair way to say it. We are not adding more prompts for this challenge.
Some of the India test places are within a few km of a training event. The split keeps rows tagged India out of training. It does not keep every nearby place out.
Here is what we actually ran. The files are public, so the counts can be checked.
The 0.390 figure is the most common training label, repeated for every test row. That label is non-event.
The script is here to checkk: github.com/ashwintemkar/opentrips/...
Outside India the file has 1,693 non-events and 971 events. The India rows in the csv are 775 events and 496 non-events, 1,271 rows: github.com/ashwintemkar/opentrips/.... The 786 in the post is the India count before some rows were dropped because the rain fetch failed. Always saying non-event scores 0.390. Always saying event scores 0.610.
0.899 is above both on this split.
The events are NASA's Global Landslide Catalog: data.nasa.gov/dataset/global-lands.... The rain totals are Open-Meteo's historical precipitation sum: open-meteo.com/en/docs/historical-.... The script that builds the table is github.com/ashwintemkar/opentrips/....
The India non-events were added on purpose. The first India test had only events, so a model could look strong by saying event every time.
add_india_negatives.pyadds random places and dates inside India so the test has both classes: github.com/ashwintemkar/opentrips/.... On those India rows, average 14-day rain is 55 mm for non-events and 245 mm for events. A simple cut on that rain column, chosen outside India, scores 0.819 on India. Logistic regression in the table is 0.821. That matches your point about rain.8 of 10 is the 10 prompts in
compare_plans.py, held out from 24 training examples. The base model passed 0. The fine-tuned model passed 8. github.com/ashwintemkar/opentrips/... and github.com/ashwintemkar/opentrips/...On the distance question, 74 of the 1,271 India test rows are within 5 km of a training event, and 130 are within 10 km. The closest is 0.04 km. The median is 99 km. 76 event rows sit inside the India box but are not tagged India, so they stayed in training: 37 with a blank country, 29 in Nepal, and a few in Bangladesh, Bhutan, Myanmar, and Pakistan.
Such a neat concept for anyone who loves spending their weekends outside! Getting a clear 'go or wait' verdict along with local trail and weather details all in one place makes planning so much easier. Awesome work👍️