Aggregated data, as defined by Wikitonary:
The state of being collected into a mass, assemblage, or sum
In web terms, this means the sum of data, as a group, taken from various feeds and be useful for later output to provide a wholesome outlook of the participant particulars.
An example of using aggregated data is taking output from a particular subject from Twitter and Sky News and grouping the data to output the latest information of what’s occurin’
Here are five tips towards creating a decent aggregated feed based on various sources. Although they may not be perfect or useful in all situations, they should provide a basic outline on what you should be doing when setting up your feed.
Tip #1
Make sure your sources are reliable! Although in some cases; i.e. taking data from social reports such as Twitter, where information given may not be 100% truthful – it may be hard/impossible to filter out the useless “WHAT’S #MAJORHAPPENING ABOUT TWEEPS?” tweets, they’d still provide an increment to how important the event may be.
Collecting your data from notably reliable sources such as:
- News websites such as BBC, Sky, ITV, Guardian etc
- Citizen reporters who have a reputation for providing reliable reports.
- Social media websites which are known for reporting true facts.
Should provide you a higher chance of providing your users with true and up to date information.
Tip #2
Stay ahead of others! Try to collect data from realtime sources such as:
- Website feeds
Tip #3
Prioritize your sources! Although I could go into how you should have algorithms setup to categorize and prioritize the information the site that it’s currently processing, people shouldn’t expect normal people to be able to come up with something like that on their own in the time of a potential crisis or several worldwide catastrophes.
Tip #4
Use feedback from your visitors! If your site is providing your users with out of date or incorrect data, then it could cause problems and confusions down the line. When this happens, you’ll start to lose visitors and potential revenue. Provide a way for your users to send you feedback on the data you’re displaying and make appropriate changes when something isn’t working as it should.
Tip #5
Know your facts! Back when China had a massive earthquake in 2008, Twitter was the first notable source to the incidents happening, but what tops the triumph off even more is the fact that the users beat the USGS with releasing the news by a pretty staggering 3 minutes. – Source and stuff.
This proves how much trusting your sources can help and knowing that Twitter was the first to release the news, means that you can pretty much guarantee that it’ll happen again and again, the more the user base grows.
This post is a bit different to my others since I’m not really writing about the happenings and what-not, it’s more of a public notepad for me to come up with ideas for my application; iTunesTweet.
The ideas that I write about may or may not become a part of the stable release of iTunesTweet. This could be due to various reasons:
- It’s not stable, i.e. it doesn’t work in the way that it was intended to or it has too many bugs for me to take care of at one time.
- There is no need for the feature. Some times I may come up with an idea that I think would be useful, however in real world testing, the majority of users would think otherwise.
- It changes a big part of the code. I don’t have time to re-write the main system again, so features will be currently limited to the size and time they’ll take to integrate.
So now that I’ve covered my ass from people who may get annoyed that a feature they really like isn’t in the application, I can begin writing ideas!
Now playing feed
This would be a global feature available on the website (once the re-design is finished). With a few modifications to the log sequence it would be possible for me to see how many songs are currently playing (ignoring the fact the user may press pause) and display them in a feed on the website.
This shouldn’t be too hard to integrate since I’d only have to work out when the song will stop playing (TTL for the song).
Random maths jibber
Current time – (song length + length played for) : something like that anyway
I think this feature would open up many new possibilities for the application. A feed of currently playing songs across the world would be a brilliant way to promote music. The more people playing the song, the more it stands out.
We could even take into consideration the rating the song has, and use this to promote it further.
Share playlist
I promise this has nothing to do with the recent movements by Spotify which now allow you to share playlists to friends at Twitter and Facebook!
Technically it should be possible for me to find out the playlist in which the currently playing song resides in. Through this it would then be possible for the user to share that playlist on their Twitter account. The website would also be able to group playlists by Twitter user (which could cause problems if the user changed their username, but it’s a risk).
It’s not a big idea nor particularly complicated but it’s a nice way to share music with the world.
Better error handling
This is a given, something that I should have implemented a while ago. Due to Twitters recent surge in problems, iTunesTweet needs to be equipped with enough decisions in its code to handle as many problems as I find.
Thanks to Cohen I was able to fix one problem this morning. And since Twitter fails so much at the moment, now is a brilliant time for me to teach iTunesTweet to handle all kinds of errors.
Eventually the error log file will consist of a deeper analysis of the problems including:
- Epoch/unix time stamps
- All data that iTunesTweet was sending
- Twitter status code
- And whatever else I deem useful to further investigation
The resulting file will automatically be sent to the server when it reaches a preset amount of errors, or you can choose to send it if the error level is bugging you.
Song correction tool
We all have songs in our libraries which have imported names such as “03 David Guetta – Memories” or with spelling mistakes “Dvd Guetta – Memorys” – maybe it’s just me, but I find them annoying and I hate correcting them as I find them.
Wouldn’t it be awesome if iTunesTweet could analyze every song in your library and automatically correct or suggest the real spellings?
I’ve applied for an API key at Song.ly in the hope to integrate this feature into the code! To reduce the amount of API calls iTunesTweet makes, it would be a good idea for my server to cache all of the original and correct strings and then run the files against the server. Once I have a base library of a few thousand songs, I could further the development of the tool to check my server before song.ly – see I play nice with API developers.
Album art for those which iTunes Store can’t find
Some times iTunes Store can’t find album art to a song because it’s based on both the artist and the song title or because iTunes just fails at these kinds of things.
It should be possible for me to write a custom album artwork handler for all of the songs that iTunes can’t handle.
Keep in mind that this feature would still be limited to how accurate the song information is, although a mix of this and the song correction tool should give you good results most of the time!
More to come
Some more ideas will probably arise because of these, or betas will come out with me testing them, so keep an eye out 
Recent Comments