Not My Usual Blog Post
Despite a career in Business Intelligence approaching twenty years, I have never used data in any meaningful way in my own life, until now: which is why I feel the need to blog about it!
I was planning a trip to Tokyo and wanted to identify the best locations in which to book hotels based on local attractions, with the intention of staying at several places over a fortnight.
However, between Tokyo being huge and the volume of entertainment, it became obvious that some sort of map was needed and the generic tourist maps available online are not personalised to my interests.
Once developed, and the hotels booked, I want to add information about places/activities of interest and fine-tune the holiday. It is too far to fly to miss anything!
Power BI is the reporting tool chosen for this task due to its easy mapping capabilities and because I happen to be doing some work with it currently.
The data is limited to the following fields:
…which is all I need at this point to identify ideal hotel locations, especially as I am manually inputting all this data.
The range of data itself is very personal to myself and is unlikely to make a quality list for anyone else.
(A good example of this is how Denny’s, McDonalds and KFC are not in the data as I’m not flying that far East to eat Western fast food! Yet, I have Starbucks listed because I bought a Starbucks mug in Kyoto many years ago and would love one from Tokyo.)
The locations which are in the data set are not all “must sees”, in fact, most are not. The data set is a list of things that may be worth considering with the ‘Favourite’ field used to identify things of particular interest.
The ‘Type’ field was quite challenging due to masses of entertainment in Tokyo being too unigue to be of any particular type! I decided to use ‘Japan’ as an ‘Other’ to contain things like cosplay Mario Carts or the Ninja Cafe!
All in all, I have over two hundred rows of data from which to work.
Map Steps in Power BI
I have not done anything remotely clever with Power BI mapping. I’ve just used the standard functionality, which is enough for the task at hand.
Here are the key points to creating the map aspects of the dashboard:
- Configure Latitude and Longitude columns.
- Insert a Map.
- Drag and Drop co-ordinates to their respective fields.
- Add a Slicer to filter and Table to show the details of a specific map item to finish off the dashboard.
At this stage, the report is just a map with a table showing the name and opening hours. Currently this is all I need.
The full view map above shows centralised clusters of attractions and a few outliers which will need evaluating as to whether they are worth the effort to see. (I expect the Ghibli Museum to justify any amount of travel!)
The following screen prints from the report show the two main clustered areas that are the prime candidates for accommodation, and an outlier that will require a dedicated trip.
Map 1: Cluster of Tourist Spots
Map 2: Another Cluster of Tourist Spots
Map 3: Outlier: Ghibli Museum
With these clusters of attractions it was a breeze to identify the ideal hotel locations and book accommodation accordingly.
A few other insights emerged, such as a lot of novelty cafes being clustered together, making the likelihood of visiting many of them unlikely, especially when there is an endless list of ‘proper’ restaurants to be visited!
With months to go before the trip, and the hotels intelligently booked, it is now a case of enriching the existing data set over time.
I need to identify which activities require booking in advance and some will doubtless not be of interest once I investigate further. For example, the aforementioned cosplay Mario Carts, looked fun at first glance, but after reading how they annoy locals, I decided against it.
A Cry for Help!!!
Currently my knowledge of ‘cool stuff’ in Tokyo is (very) limited by what I can find on Google. With that in mind, anyone out there with any suggested activities and/or eateries in Tokyo, please leave a comment on this blog post!