Thursday, 9 October 2014

Finer details

Which social networks API's?

I've thought about which social networks I can use and have done a little research into the APi's.

I started off by looking at Facebook's API to see if it provided what I wanted. In one way it did but overal It didn't. Although Facebook has trending topics there isn't a search function, the only data you get when you pull the data is that of your friends that you are connect to or pages you've liked, which isn't what I want.

For this reason Facebook will not be used.

The Twitter API was is much more useful there API allows access to all location based tweets, so long as users have location based services enabled.

Instagram allows for location based services also, granted it's harder to tell with Instagram. People still hashtag on there I will see if I can use the hashtags people create and collect that data, and put it into a google map.

Which languages will I need to know?

Advanced knowledge of HTML5, CSS3 will be required. I will be using a CSS pre-processor called SASS as my styling tool. I'm also toying with the idea of using Angluar.js but I need to research that more and discover what it does.

JavaScript will be use as the main programming language but I may have use ruby/python just to install some extra's, but this won't be difficult. I want to use node.js as a pure JavaScript backend server, just to store past data on.

jQuery will be used throughout this website, not just in the design of it but in the way that I use JavaScript. I will be using jQuery AJAX calls to collect data from API's.

I want to use D3.js to create my own graph, which will be fully interactive. I am told that D3.js is a steep learning curve but very worthwhile. D3.js stands for Data Driven Documents. These documents are entirely interactive.

What research is required?

The main thing for me to achieve this well is to make sure I research how to use D3.js properly. I will also need to look into the correct way of data mining to ensure I get the correct data I need.

When I am researching design I will look at old designs and old examples of my work. I will see how they do it and take notes. I will then look at new designs within web, I need this to be responsive.

I have to look at what I know about HTML5/CSS3 currently also, and maybe advanced my knowledge of that. However I am pretty confident. (Except HTML5 Canvas)

Github will be used as a code repository, so If i want to experiment with my code and completly mess it up Github will allow me to revert back to an older working version.

Idea

Example:

If an agency was wondering when the best time to put their promotion on television was they would be able to use this tool to market there target audience.

For example if there product was a new type of Alcohol and they needed to establish what days and times most people are using social networks to talk about using alcohol for entertainment purposes. They load up this website and see that on a Friday in San Francisco around 18:00 that a lot of people are talking about leaving work and going out drinking for the evening. They can then plan to book airtime at the appropriate times.

The agency would be able to look at passed data also, which would show how well products previously done.

Details:

The website itself isn’t just aimed at the alcohol industry it will be multipurpose and will be used by pretty much anybody who is looking to market a product via Television.

Another great thing about this idea is that over time it collects a range of data from previous days and stores this data. Which can then be represented via a heat map. Current and in the past

This allows advertisers a method of looking at trends, from this they will be able to tell what times are best for there current and future products.

I should mention the way the heat map is going to work. When people send tweets/post pictures or update there status most of the time they leave there Geolocation turned on. Using social media API’s you can manipulate this data into a Google map, which you can then manipulate again to turn into a heatmap.


From this advertisers would be able to use this tool to work out which time suits them best.