ASWI - Pokročilé softwarové inženýrství » ASWI 2021 » OpenData vlastní téma (KIV) - Tři Mušketýři

h1. Project details

h2. Data

The application uses datasets containing historical information about the weather on campus and activity based on JIS card verifications and historical WebAuth data

* JIS data - data about JIS interaction on ZČU (source: http://opendata.zcu.cz/Snimace-JIS.html )

* Weather data - historical weather data on ZČU campus (source: http://opendata.zcu.cz/Energeticky-dispecink.html )

* Authentication system - data containing history of log-ins to school computers on ZČU (source: http://opendata.zcu.cz/Autentizacni-system.html )

Data is processed in regards to buildings, and model can then combine those information sets in a way that produces the most sensible results. More details about data is written on page [[Data sources]].

h3. Weather

Weather data for model training contains the following information: date and time (dd:mm:yyyy hh:mm:ss), temperature in °C, wind m/s, rain (0/1), light values in lux. Weather condition values sunny/partly cloudy/overcast/dark are approximated from data input in lux values.

User can input the following values:

* date (dd:mm:yyyy hh:mm)

* temperature in °C

* wind in m/s

* probability of rain (0-100%)

* weather conditions (sunny/cloudy/overcast/dark)

However from training data only data from 7am to 6pm will be included in training. That is because this time period is deemed relevant to university class attendance.

User has two options: use automatic weather prediction or input all of those values manually, if they choose so. Automatic weather data is downloaded from server http://wttr.in/Plzen,czechia?format=j1

This data source is a JSON data file and contains data from today, tommorrow and the day after tommorrow, all in 3h intervals between 0am and 9pm.

h3. Datasets

Datasets are freely availible online, and every month should be updated with new dataset. Adding new data to already running application needs re-training of the model, so the responsibility to trigger download and retrain of model belongs to a server administrator and can be triggered any time.

There will be possibility to include in retraining only a certain part of input data, based on requested time period.

h2. Design

In this chapter is the general description of application capabilities. Details about architecture and communication protocols can be found in [[Project architecture]]

h3. Client application

Design propositions can be seen on page [[???]]

Output from application is a heatmap and percentage of predicted attendance (0-100%).

Requests that are sent to server are the weather values specified in chapter Data - weather.

h3. Server application

Server needs to provide the following options for server administrator:

* retrain model

* download data

Furthermore server needs to process and answer client requests. Server answers with percentage of activity (aka how much traffic is predicted to be in a specific building at a specific  time taking into consideration specific weather conditions).

That means that server application needs to interract with client, server administrator and with online data sources.

h2. Implementation

h3. Client

The application is a mobile app, or WebGL app. It is developed in Unity which provides more possibilities for export with modifications for given platforms.

* Mobile app

Android support 5.0 and higher, use Unity UI tools to provide touch support.

Deliver an apk of the application to download outside of the Playstore or optionally, create a developer account.

* WebGL app

This provides the option to embed a Unity Canvas app in a simple website. Possible issues with support in various browsers, but unlike the mobile app, the university could provide hosting.

h3. Model

Model will be based on multiple Naive Bayes classifiers.