Exploring Places and Area Region of Surabaya
Introduction
Surabaya is a big city of Indonesia. Instead of Jakarta, capital cities of Indonesia, Surabaya are second biggest cities in Indonesia. With big city, of course we should expect high about the developing program. With 326.81 km squared of areas region of Surabaya, actually we should hoping that building development is quite fast [1].
Since there are so many building we can develop, such as restaurant, coffee shops, supermarket, etc that can helping for financial of Surabaya, because with growing businesses, of course many people would get a job, and decrease people who don’t have job.
My problem is how we can know which neighborhood that is very crowded with building, and which is less crowded, so that we can create more and more building.
Data
Based on problem above, we need so much data about Surabaya itself and Foursquare API to helping us know what is nearby that area. So data sources we needed is:
- Data of areas region of every neighborhood on Surabaya
- Data of latitude and longitude of every neighborhood on Surabaya[2]
- Data of Foursquare API for understanding what building in radius of every neighborhood on Surabaya[3]
Methodology
In this project, we want to know what kind of places near of every neighborhood, so that because Surabaya is so big, I decide to create radius 1 km for every location, to make sure that everything is gathered.
After that I want to know how many places are gathered, and can create some visualization of it, and then linked it with areas region, it is quite dense or not.
Then, using K-Means to helping us understand actually it can diversed into how many cluster to understand that maybe we can build something on this neighborhood, just like that.
Analysis
First, I am analyzing data with Folium[4] to understand what kind of maps that we want to take a look.
After that, I want to analyze data of Area Region of Neighborhood in Surabaya.
From this data we know that majority of neighborhood on Surabaya has Area Region between approx. 7 until approx. 13 km squared, while there are some outlier outside 20 km squared.
Before we go further, let’s take a look of our data.
The data consists Neighborhood, and then Latitude Longitude, most important things is Venue and Venue Category to know what is in area of neighborhood itself.
And then analyze venue near the data, for now i am using LIMIT equals 200 and radius is 1 km since that Surabaya is a big city. And then get the data venues.
Woah, that’s so much things in Genteng and Gubeng. Let’s check what is the most venues on the neighborhood.
With 8 of 31 Neighborhoods, Indonesian Restaurant is dominated what is the most common venue on Surabaya.
Let’s check how many cluster we can create with K-Means. If you don’t know K-Means, there are so many introduction of it. You can check it here
Since it is hard to make what is good k-, so I decided to 3 cluster, because if I choose 8, there are so much cluster and maybe it’s hard to interpret.
Last but not least, merge all of data we have, using 3 cluster, create maps visualization that show area region and cluster it self. And we get like this.
Result and Discussion
After so much time consuming, finally we can get the result. Our analysis show that it’s great to know that with how much your area is, it is still doesn’t have any impact, except for Indonesian Restaurant that they are spread out to 8 in 31 neighborhood to make it top places in Surabaya.
And then we absolutely know that in Genteng and Tegalsari is very crowded with just 0,04 km squared per each venue. While only 12 from 31 has density above 1 km squared. In some neighborhood is so dense, some of them not.
Conclusion
So after looking at the maps above and the data itself. It is clear that Surabaya is dense, when you sample just 2 or more neighborhood, such as Tegalsari and Genteng. But if you see in depth, there are some land that we can build something. If you want to start your businesses, go ahead for Indonesian Restaurant. You can build something that can improve economy of each neighborhood. Maybe in some places there are few venues gathered because the data is not complete or what kind of else problem. But in this I assume that we need more venues, but not in dense neighborhood. Take a look at Lakarsantri, Rungkut, Sukolilo, and Benowo.
Well I know that this project is still potential to grow, especially with the data itself is growing and you can use your city maybe, if you have enough data of it. For full explanations and code, take a look of my Github here.
This is taken for my Linkedin articles, connect with me here.
See you on the next post!
My Reference
