COVID-19: Neshan users can now monitor crowded and heavy pedestrian traffic areas in Iran
In the new version of Neshan, we have added a layer that displays heavy pedestrian traffic areas. However, staying home is the best health measure these days, in case you have to leave the house, you can use this layer to monitor crowded areas and avoid them. All you have to do is click on layers and choose “heavy pedestrian traffic areas (Specific for COVID-19)”.
As you can see in the image, you can select the layer and see those areas in the last 24 hours with purple color. Also, orange and yellow areas are the current heavy pedestrian traffic areas. Orange has more traffic than yellow!
In the following, I want to explain in a simple way how we found these places.
How it all started?
With the rise of Corona cases in the country and reports of the first cases in Mashhad, all Neshan teams started to work from home. After a day or two, the number of cases went very high and Javad, our manager, asked for an online meeting. In that meeting, all plans for Nowruz were suspended and we were urged to find ways to help the public in the current situation.
Following a thorough evaluation of other countries’ actions against the epidemic, we noticed special effective methods in eastern Asia countries such as China and that was highlighting the high-risk areas that people could avoid. So we communicated with Baidu, the giant internet company in china, and asked them to share their experience about what they had done. They confirmed that showing the possibly infected areas on the map effective in China. Accordingly, We decided to do the same but couldn’t find a valid source of data for that. We then concluded to use our own data and extract what can help here. Meanwhile, we were aware that time is running and we should do something about it. Finally, the decision was to evaluate some ideas and work them out at the same time.
The idea of working on heavy pedestrian traffic areas was assigned to our team, which is generally in charge of data analysis in Neshan. The idea of engaging users for contribution was assigned to the crowdsourcing team. A group of product owners went to follow special ideas. We had the speed and location of users who were using the app in addition to the start point and destination of the users’ navigation, a large database of POIs and maps of streets and neighborhoods.
We got to action
As mentioned above, we used our current data to highlight the heavy pedestrian traffic areas. Since our team is also in charge of traffic prediction and time estimation, we have a substantial amount of data analysis experience. But this was different in the sense that we had to analyze pedestrians not vehicles! This was not what Neshan was initially built for.
The first thing that came to us was using low-speed data which is ignored in the traffic estimation. For example, when users are not driving. They are sitting or walking and also exploring the map. Their speed is considered as almost zero here. This sort of data is important to avoid or our traffic estimate will go very wrong.
The proportion of this data to the total is not very high, about 1 to 2 percent. It makes the data analysis easier but those who work with such data know it’s not good at all. Then, Ahmad suggested user contributions. Neshan users can report the heavy pedestrian traffic areas in the latest version of the app.
Data and reports combination took a lot of time and we finally managed to find a way to organize them. The results were mixed, our examination showed, although most of the area we displayed actually had heavy pedestrian traffic, some of them were wrongly highlighted.We also decided to add the requests for travel data (source and destination chosen for travel) in this output. That is to say, those locations who were chosen the most are more likely to have heavy traffic. Also, POIs can assist in the determination of heavy pedestrian traffic areas. These two methods helped us modulate the ultimate output and optimize it.
The biggest issue in projects based on data analysis is to convert them into a service which can be continuously updated to generate appropriate outputs based on the latest data. In this case, this means to update pedestrian traffic on an hourly basis so that the service is more practical.
The data should finally be displayed in a suitable way. Despite our long time experience in this, we had to create new algorithms to display the data better which was eventually satisfactory.
The first release
The current features on the map are the result of 24/7 progress in the last two weeks. This is the very first release and our field examination of it in Tehran and Mashhad yielded satisfactory results. Nevertheless, people follow health instructions and there are fewer heavy traffic regions.
Our team has a lot of ideas for this project. For instance, determination of potential areas of heavy traffic through evaluation of past data. Search data can also be used to determine more accurate data. The user contributions are rising daily and future versions will be more effective.
Our aim is to release a new version with more features in the coming days. If you have ideas that can help us provide a better service and improve the current conditions, please share them with us. We hope to be able to serve better and together pass this difficult time.
You can download the latest version of Neshan for android here
At the end, we would like to sincerely appreciate the Baidu team sharing their valuable experience of fighting against COVID-19, which helped us put our ideas into practice. COVID-19 has once again challenged us to unite to protect human life, as it is, we, in Neshan, would be more than happy to share our experience of calculating pedestrian traffic or work in cooperation with other teams to implement IT-based solutions to the management of the pandemic.
