In this section we will apply modeling techniques to infer more advanced insights. We have already seen how space time and other factors such as age and type of vehicle affect traffic accidents. This is however something to be considered in other cities as it might have a big impact on accidents especially in cities with more extreme and variable weather conditions.ĭiscovering Complex Data Patterns using Advanced Spatial Modeling Surprisingly enough we found out that in Barcelona the difference in the number of accidents during rainy days vs non-rainy days is not statistically significant most likely due to the city's mild weather. We were also interested in understanding the impact of weather and in particular the impact of rain. These insights further underline the importance of defining dynamic hotspots. Almost 50% of the deadliest accidents occurred on a weekend.Car-pedestrian crashes are the main type of accidents for people aged +65 and the fourth when considering all age groups.While young adults aged 18-25 are the fourth highest demographic group in number of accidents they are the second when considering only accidents at night.Half of the accidents with alcohol as the direct cause occurred on weekends.Some interesting insights we can easily identify are: This data is later transformed into Vodafone's 250x250m cell grid as described below. Working population dataset is provided by Unica360 and consists of the number of companies and employees by type in a 100x100m cell grid.We are using a month of data because we are interested in identifying variations between days of the week and time intervals not exact numbers. Human Mobility (footfall) dataset is provided by Vodafone and consists of anonymized counts of unique visitors and total visits to an area during a time window segmented by age gender visitor profile and economic level in a 250x250m cell grid.POIs are classified at different category levels the highest being by trade division (retail transportation tourism etc.). Point of Interest (POI) dataset is provided by Pitney Bowes. We are using a month of data because we are interested in trends not exact numbers so one month is enough to identify hourly and daily trends.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |