Archives for 2014

publicartio:

Im going through the PA database today and removing all the images that are not geotagged. This is the first round of getting the best set of data for analyzing geotagged graffiti.

Almost 5/8ths of the 300k images stored were not geotagged. I didn’t expect the percentage to be so high. By removing these images, I’ll have a better idea of the actual number of query-able images.

Charts coming soon.

Im going through the PA database today and removing all the images that are not geotagged. This is the first round of getting the best set of data for analyzing geotagged graffiti.

Almost 5/8ths of the 300k images stored were not geotagged. I didn’t expect the percentage to be so high. By removing these images, I’ll have a better idea of the actual number of query-able images.

Charts coming soon.

After releasing the Public Art, I felt like I wanted to walk away from the project. I was mentally drained and didnt enjoy the topic anymore. I started looking at a service to help groups (friends/random) coordinate activities together. Its become more of an experiment than anything. Now that I have settled on the new project, I have gotten back to Public Art.

A few major things that were missing from the release are now on my plate. First of all, the real time feed to Instagram has been paused for months. After the first two weeks of data collection, I paused my image collection process. I was aware of the fact that I was collecting too many images too quickly. I also was running out of database space and knew there was no inherit value for collecting so many images. Now I have gotten the the point where I am flexible enough to address the latter.

I realized that I completely screwed up the initial data set up. I was using various ruby applications that manually patched together database content. Im moving over to bringing everything under one roof and setting up the entire web/api section under one application. In the process, Im reintegrating the data polling to grab Instagram pictures again. This should be running by the end of today.

I launched the iOS app with no extended features. The sharing tool wasa ghetto email link with bad data passed. The flag tool didnt work. It wasn’t even being properly tracked. Im getting to that now. Im creating a system for flagging to work.

I still have a few parts I need to think through. Now that the image polling is reactivated, I need to do something with all this “stuff”. In one of my previous experiments, I was creating individual webpages for each crawled image. This let me get my long tail Google game up. Im going to do this. The only question is how to do it effectively.

I haven’t done anything interesting with the tags. The tags could potentially reveal a lot. I expect there are some tag clusters that would be interesting to follow. I might even be able to do some heavy quality control if I can understand the tag-to-quality relationship.

There is no favoriting system yet. This might not be necessary in the long run, but if the app is to be useful, I might want to create a way for batching/lists. The process of browsing the database makes it very difficult to track down old results. I might create a list of old searches. I might also create a “recently viewed” list.

All this stuff takes time, effort, and attention. If anyone is interested on working with me on this, please reach out! Im looking for front end developers, rails develops, ios developers.

I often think my projects will last forever. Or at least until I stop paying the domain service or the hosting service goes kaput. I have yet to have my domain payment outrun my hosting service. I think this is common for most people.

We build with the notion that our projects will live on. This is why advertising projects or marketing projects can be soul crushing. The work you invest yourself into is predetermined to be short lived.

I want to make some projects that are purposely timed in their value. Specifically, I am fascinated by the viral growth a website or app can develop through their service growth. What if this viral growth wasn’t sought to be sustained. What if the viral growth had a time limit.

What would a service look like if it exploited the variables associated to viral growth, while simultaneously engineered to collapse.

I assume the average viral video starts out with a slow natural growth or lack of. Once the factors that influence the video’s sharing emerge (through blogs, social networks, accidental link sharing, etc) the videos growth rate explodes. Post-explosion and post-hype, the video returns back to its relatively nonexistent view rate. This results from the natural process in which the ‘new hotness’ overtakes the communal attention span. Ads, youtube videos, hot single releases all get this viral growth curve.

The eventual decline is often undesired. The factors that determine the shared contents viral elasticity are dependent on the sharing communities. These communities are influenced mainly by their exposure to the content and their relationship to the people they would want to share with.

I assume all of the factors involved could be manipulated to some extent. In store ‘flash-sales’, importance is placed on time by providing a deadline for action. In social networks, the ‘aha moment’ often results from a certain tipping point where a certain number of a user’s friends are online. The factors that influence the likelihood for buying or sharing or joining a service are readily manipulatable, so a service that purposely seeks to get critical mass isn’t new.

Often by mistake, the services that get critical mass have some ‘thing’ that results in a leaky bucket syndrome. People stop discovering the service, start leaving, stop returning. Its often an uncontrolled phenomena.

Instead of this eventual fatality happening, what would a community or online service look like that depended on this.

Edit: From Tina – Self-destructive game – dominikjohann.de/impetus/

https://blog.rememberlenny.com/2014/12/28/i-often-think-my-projects-will-last-forever-or-at/