Real-time delivery route optimization

1) Context and problem

In logistics, even small delays along a route can lead to missed delivery windows, higher costs, and a poorer customer experience. Routes are often planned in advance and do not adapt to dynamic conditions such as traffic jams, weather changes, or local incidents. Dispatchers see only a partial picture — where the courier is now, how many stops remain, and how the traffic conditions are evolving. SMART LOGISTICS should bridge this gap by predicting the probability of delays at each route segment and suggesting optimizations: reordering stops, adjusting departure times, changing the route, or offering detour options.

2) Data provided

3) Expected product

A service that:

• predicts the probability of delays along a route
• suggests an optimized order of stops
• recalculates departure time and optimal route based on real-time conditions
• provides recommendations in a dispatcher-friendly format

4) Constraints

5) Success metrics (aligned with evaluation criteria)

6) Final deliverables

Participants must provide three mandatory deliverables:

1. repository access. Sufficient technical information required to validate the solution: frontend, API snippet, Postman workspace, Docker/Python build test, or any format that allows experts to run and evaluate the project
2. link to the project presentation. Explaining the idea, system architecture, delay prediction method, and example optimizations
3. link to a working MVP or clickable prototype, demonstrating the key scenario: delay prediction and route optimization suggestions

When the bus will actually arrive

1) Context and problem

Urban transit apps usually display scheduled times but do not account for real conditions — traffic, delays, bus speed, or weather (rain, snow, wind). A user sees when the bus is supposed to arrive, but not when it will actually arrive. This leads to long waits and missed connections. PREDICTIVE TRANSIT should address this challenge, transforming transport and weather data into accurate and clear short-term predictions.

2) Data provided

3) Expected product

A service that:

• predicts bus arrival times under real conditions
• estimates how many people are waiting at a stop
• displays the forecast in a convenient visual format (map, card, timer)

4) Constraints

5) Success metrics (aligned with evaluation criteria)

6) Final deliverables

Participants must provide:

1. repository access. Frontend, API snippet, Postman workspace, Docker/Python build test, or any runnable format
2. link to the project presentation. Explaining the idea, architecture, prediction method, and evaluation results
3. link to a working MVP or clickable prototype. Demonstrating the main scenario: arrival prediction and stop occupancy estimation

Weather with actionable advice

1) Context and problem

Weather apps show temperature, wind, pressure, and precipitation, but they do not explain what this information means for the user right now. A person sees the numbers but does not get an actionable answer: Will it be comfortable in the evening? Should they reschedule an outdoor meeting? Will the wind be dangerous? Do they need an umbrella? As a result, forecasts remain background information — informative but not useful. WEATHERWISE should bridge this gap, converting raw weather parameters into short, relevant, real-time suggestions.

2) Data provided

3) Expected product

 A service that:

• converts weather data into human-friendly recommendations
• generates short textual tips
• works as a mini-app, widget, or AI assistant inside a social network remains understandable even when the data is incomplete

4) Constraints

5) Success metrics (aligned with evaluation criteria)

6) Final deliverables

Participants must submit:

1. repository access. Any format allowing experts to test the solution (frontend, API snippet, Postman workspace, Docker/Python build)
2. link to the project presentation, explaining the idea, logic, architecture, and results
3. link to a working MVP or clickable prototype, demonstrating the core user scenario