Dam-It

Inspiration

The Brumadinho dam disaster occurred on 25 January 2019 when Dam I, a tailings dam at the Córrego do Feijão iron ore mine, 9 kilometres (5.6 mi) east of Brumadinho, Minas Gerais, Brazil, suffered a catastrophic failure. The dam released a mudflow that advanced through the mine's offices, including a cafeteria during lunchtime, along with houses, farms, inns and roads downstream. 270 people died as a result of the collapse, of whom 259 were officially confirmed dead, in January 2019, and 11 others reported as missing, whose bodies had not been found.

If people were more aware of the safety measures to be taken during such disasters, the death toll could have been reduced. So we came up with an idea to educate people about dam collapses in general. Since everybody enjoys playing games, we thought that creating a game would be an engaging activity to spread some life-saving information.

What it does

The game creates scenarios using real world data. It is a turn based strategy game, where before a player's turn he/she needs to decide on a number of actions to take to score-points. The actions are as follows:

1. Choose which settlement to evacuate (only a limited sites can be evacuated)
2. Choose where to relocate the people (guess the safe sites)
3. Some options to take preventive measure. Options will be provided based on each level - player needs to choose one among the many provided in each level
4. Preventive options will be like whether or not to go about cutting trees in the surrounding area, construction of roads, settlements, number of frequency to inspect the dam, etc. each of this actions will have impact of whether the dam collapses or not.

How we built it

We have used Unity Game engine to develop the game. We used a real world data consisting of about 90,000 dams and used IBM-Watson studio to estimate which features are responsible for a dam to be hazardous. We had planned to use google-maps-api to extract the elevation information at each of these locations and generate a realistic level where people can try to prevent the damage to the minimum

Challenges we ran into

The Watson studio generated a classifier and according to it the most important feature for dam to be hazardous was its inspection frequency. So clearly there is an observer bias in the dataset. We plan to use some more datasets to improve our predictions.

Accomplishments that we're proud of

We are proud of being able to use the IBM-watson studio and analyse our data within a short period of time. We are proud of collaborating with like minded people, and learning a lot from each other.

What we learned

We learned how to infer results based on a linear regression model. We learned about gamification can be effective for educational purposes.

What's next for Dam-It

Create better models for predicting the hazardous nature of dams. So that we can inform the government policy makers about preventive measures to be taken

Built With

• callforcode
• google-maps
• ibm-watson
• unity