JoeIsDone - North Carolina map LIVE

JoeIsDone ^

[byecomey]

Click through to the North Carolina map. It is a different flavor than the Florida map - this one is not for live updating so much, it's for browsing and playing with demographic data.

Note that the 'database' is entirely client side (compressed from 5GB to 4MB). It seems to work OK on my cell phone but a little lag is expected in the initial load and again on the queries.

[byecomey]

Ping to https://joeisdone.github.io/northcarolina/ . It is not the addictive factor that the FL map is, but it will let you browse and play with demographics. The split view is “Party Advantage.”

Lazamataz - I hope this shows you what I was going for.

I’m exhausted. I will add in the totals table a bit later.

[Chgogal]

Bookmark

[thinden]

Neat tool

Great job

Thanx for sharing

[Ravi]

Great job!

[LS]

If you aren’t following byecomey’s interactive HOURLY updated maps of FL and NC, you’re missing a great exciting time.

[Jane Long]

Thank you, byecomey, for this great tool/resource.

Thanks, Larry, for the Good News ping.

[Guenevere]

Bump

[laplata]

Thanks, Larry!

[Lazamataz]

Hey byecomey, now I understand what you are doing, having looked at the maps. Unsure if SQL would help here, so your direction looks right.

I’d be interested in having a voice communication to understand your architecture. Always willing to learn a new way to do things!

[Lazamataz]

It’s a different approach to have the data on the client side. Interesting. Not dangerous in terms of PII or PFI, so that’s good, but wondering if any security flaws need to be addressed. In what fashion is the db sent to the client? Blobs have some security concerns, for example.

[byecomey]

Am impressed with your distributed systems expertise and would like to discuss this topic with you as well. Shoot me an email at the contact info on the site whenever you’re up for it.

[Irish Eyes]

For later

[byecomey]

Ah, okay, so the DB is public information. The original CSV file is over 5 GB file, I compressed it to 4MB by encoding the important information on the rows and mapping these to aggregations.

This is the "database". The keys are rows. The values are aggregations (counts of each row). So, for example, if you see the entry "1079847292": 5, that could mean "there are 5 women of age 65+ living in Wake County..."

Here is how the decoding is done on the client end - Javascript file here.

This kind of bit packing and bitwise matching is technically how some business intelligence database achieve dense multidimensional indexing on the kernel end. After all, each new index is technically like adding a mini table to the database. At some point it gets counterproductive to have an index table on everything.

[Lazamataz]

We shall talk my friend. :)

[perfect_rovian_storm]

It’s amazing work, byecomey. Thank you!

It’s pretty laggy on my end. Might just be heavy traffic from Larry posting about it on Twitter. When you click in a checkbox, you’re not sure whether it’s going to do it or not, then it does. Having the data on the client side would probably be beneficial in cutting down on that.

[perfect_rovian_storm]

Duh...meant having the data on the server side! lol

[byecomey]

Yeah, there’s an optimization I need to do which will help with the lag somewhat. I’ll do it on the lunch break.

[Alas Babylon!]

I just checked North Carolina.

It says: 2.08% Advantage Biden

I'm NOT excited by that, Larry.

[God luvs America]

brother- i hope you are right about NC!!