MAT-Builder datasets

<p>The archive contains two datasets that have been used to empirically evaluate MAT-Builder, a system to generate multiple aspect trajectories.</p> <ul> <li>The first one is located in the &quot;rome&quot; folder and contains 26395 trajectories from 3181 individuals.&nbsp;The trajectories move over the city of Rome and were collected from OpenStreetMap. The folder contains also auxiliary datasets, i.e., the set of POIs within the province of Rome&#39;s boundaries (downloaded from OpenStreetMap) (see the &quot;poi&quot; subfolder),&nbsp;historical weather information (downloaded from Meteostat <a href="https://meteostat.net/it/">https://meteostat.net/it/</a>) (see the &quot;weather&quot; subfolder), and a dataset of social media posts from the individuals which was generated synthetically (see the &quot;tweets&quot; subfolder). All the datasets are pandas dataframes, except for the POI dataset which is a ```geopandas``` DataFrame.&nbsp;All the datasets have been stored according to the parquet format.<br> &nbsp;</li> <li>The second one is located in the &quot;geolife&quot; folder, and contains the [GeoLife](https://www.microsoft.com/en-us/research/publication/geolife-gps-trajectory-dataset-user-guide/) dataset.&nbsp;The dataset contains 17621 trajectories from 178 users. The timestamps of the trajectory samples have been adjusted&nbsp;from the GMT to the GMT+8 timezone. As in the former dataset&#39;s case, this folder contains also a dataset of POIs, a dataset of historical weather information, and a dataset of social media posts that were generated synthetically.</li> </ul> <p>For more information on the MAT-Builder project (i.e., published papers, how to use to datasets, how the information within the datasets is structured, and so on) we refer to the MAT-Builder&#39;s GitHub page: <a href="https://github.com/chiarap2/MAT_Builder">https://github.com/chiarap2/MAT_Builder</a>.</p>