MIT App Inventor Gallery

Drought_Predictor_App

July 31, 2022, 3:58 a.m. Likes: 1

The Drought Predictor App accepts the user’s zip code and a date and provides information about drought levels for that area and date. The app can also predict drought levels for a date in the future using a machine learning model and provide preventative measures. The app suggests actionable strategies that the user can follow to reduce drought and motivates the user to continue saving water. My app was designed with the user in mind and allows easy navigation from screen to screen. The logo in the opening screen is a tap with a single drop of water falling from it. This symbolizes that every drop of water counts and every proactive step taken by us to conserve water matters.

Credit: 

https://droughtmonitor.unl.edu/

https://www.pngwing.com/en/free-png-dhoqh

https://www .istockphoto.com/vector/mountains-lake-and-river-landscape-silhouette-tree-horizon-landscape-wallpape r-gm1206973564-348304406

https://www.pinterest.com/pin/blue-home-page-icon-png-icon-png-button- home-png-image-with-transparent-background-png-free-png-images--842736149024233445/

https://vis ibleearth.nasa.gov/images/89110/drought-continues-to-grip-southern-california/89113w

https://ww w.explorium.ai/blog/the-complete-guide-to-decision-trees/

https://www.noaa.gov/

https://p lay.google.com/store/apps/details?id=com.discs.citizenscience&hl=en_US&gl=US

https://ww w.drought.gov/states/california

More Info: 

The home screen contains links to three screens: the About screen, the Predictor screen, and the Action screen. The About screen contains information about the functionality of the app and how it processes the information obtained from the user. The Predictor screen asks the user to input their zip code and a date, processes these values by passing them to a machine learning model and displays the predicted drought severity level in a visually engaging manner. The Action screen provides a friendly checklist of effective strategies that the user could follow to instantly save water. It also provides feedback to encourage the user to continue conserving water. To protect the user’s privacy, no data about them is stored in the app.

The Drought Predictor App incorporates a machine learning model to predict the drought severity of the user’s location based on the zip code and date they entered. The data is obtained from the USDM and National Oceanic and Atmospheric Administration (NOAA) website. Drought severity is classified according to USDM’s categories of drought severity in increasing order: D0, D1, D2, D3, and D4 with D0 representing dry condition and D4 representing exceptional drought. After preprocessing the data and finding the correlation between several factors that influence drought, many machine learning algorithms were tested. After several experiments, I found that the model that best fit my requirements is the decision tree classifier algorithm, a supervised machine learning model that uses a set of rules to make decisions. This model is about 90% accurate in predicting drought and classifying it according to the USDM's levels of drought severity.

Load App Into MIT App Inventor
Other projects by same author
Report Project
Permanent link:  https://gallery.appinventor.mit.edu/?galleryid=ffd86607-c922-44fc-af5f-d9d7e5010960