This week there were several useful additions to Hive Analytics, an app for analysing the Hive Blockchain powered by data from #hivesql
https://eroche.shinyapps.io/hiveanalytics/
Updates This week
Tags
There are two places where you can now analyse Tags
On the Activity Tab you can search for a tag and see the posts that were made within a date range.
Thanks @demotruck for this suggestion
On the Tags Tab you can see a breakdown of the tags used for the last 7 days
User Map
The highlight this week is the interactive user map. This parses the location information on the full set of user accounts and where a location has been specified by the account holder.
- It's interesting to see the trends over time for some of the countries in the table below the map.
Try the App out today and let me know how you find it in the comments!
https://eroche.shinyapps.io/hiveanalytics/
The location data is very patchy and tricky to work with as people will put all sorts of variations. UK can mean Ukraine in some cases. I have been running some queries to find Brits, but it's often not obvious where they are. Have to use other clues that are harder to automate for. Still cool to have that map. Nobody in Greenland?
!PIZZA
Really good point on the Countries. There are a few wiseguys in there too living on the Moon :)
The algorithm I am using has been tweaked some time back, but I will keep working on it to improve it. Adding in a check on languages used in posts and a few other tricks might improve performance.
You can derive some insight from Pinmapple, but a lot of that is people travelling around. It is useful to have some location data and we can see some hotspots like Venezuela. At least we have some useful tools to do analysis with. @HiveSQL is great.
Pinmapple is really cool.
btw I did find 11 accounts in Greenland I was missing :)
Wow. They may still have the lowest density of Hivers.
@eroche! I sent you a slice of $PIZZA on behalf of @steevc.
Learn more about $PIZZA Token at hive.pizza (2/10)
Does it track location solely based on what is specified in the user profile info or is it like a super computer that will also scour through the user's content over a period of time to make a super smart bot decision?
At the moment it's just looking at the user profile info and with some clever text matching it's giving really good results. It can always be made better
It should be possible to extend this incorporate other information also. Some ideas I would like to try out are
There are some fairly sophisticated GEO matching algorithims out there (for a fee) and where I am seeing stuff that the other methods dont perform well on I may look to using some of these if feasible. Better to start with the easier stuff first if its works good enough :)
Hi @eroche how do you do my friend :)
Its so Long time not see your blog buddy
Hi buddy. Its been a while, good to see you.
It's great to see you again @eroche, how are you there, is the pandemic still there. In Indonesia too, even now in the province of Aceh is improving health procedures in every community activity.
Good to hear. The pandemic has affected everyone. In Ireland we have been in a pretty heavy lockdown for most of the last year. We are doing well now and the country is beginning to emerge in the last few weeks. How is it in Aceh, Is there light at the end of the tunnel?
Glad to hear good news from you @eroche, In Indonesia, especially in Aceh the government still enforces a curfew for the community, community activities are limited to 10 pm.
No more time for coffee as usual here. Washing hands, wearing masks, and keeping a distance are still health protocols.
We are adapting a new habit here, there is no lockdown, only limiting the space for people to move.
And I think it's boring enough, we're still like last year. I don't know when this pandemic will end, since the deaths from Covid19 in India increased last month, our country has started implementing health protocols such us last year.
Geographical data might be rough but it mirrors tendencies seen among posts, strong Venezuela, Nigeria, Indonesia, I believe a lot of marketing efforts directed there could give outstanding results in the long term. We need either one killer app or one country mass adopting Hive.
I like this map of users, thanks for sharing