People usually post their life purchases on Instagram - be it fancy coffee, milk tea, cakes, or other note-worthy items. Businesses ask their customers for feedback through survey links usually printed in their receipts. Why not hit two birds with one stone and make Instagram posts their survey response.

What it does

  1. Instead of having customers fill-up online surveys, businesses can use the Survey Tracker filter to encourage their customers to rate their items, such as drinks, and post it on Instagram with the rating they gave and the filter they chose.
  2. The Instagram filter detects any partnered brand’s logo.
  3. The customer will be asked to rate their purchase.
  4. A discount voucher may then be shown to incentivise future usage of the filter.
  5. The customer may then select from an assortment of filters depending on the brand detected.
  6. After taking a picture or video of the finished project, the customer may now post it on Instagram.

How we built it

  1. Using a Target Tracker, we detected the appearance of specific defined patterns thus satisfying our need to detect varying logos. We provided a sample logo in the Image Gallery below which can easily be edited later on depending on the need.
  2. A tap listener was used to detect the given user rating.
  3. An animation was used to display/hide a coupon, should there be. We then used its Native UI Picker to allow the users to browse through the collection of effects defined by the logo detected previously.
  4. When an effect is picked, the script reads from a configuration file the changes needed to do to the scene to produce the desired effect.

Challenges we ran into

  • We had to ensure that the effects and the brands that they were tied to are easily modifiable. As such, effects had to be created as either configuration or as a singular material to be used by the script .
  • The Target Tracker only allows the tracking of one image thus limiting the number of tracked logos to our provided image.
  • Networking is currently disabled in Spark AR, which in turn, completely disables any interaction to a server or other clients.

Accomplishments that we're proud of

  • The team is proud of being able to create an interactive filter which combines the multiple capabilities provided by Spark AR. We are able to make use of the Target Tracker and the Native UI Picker to create a survey filter.
  • We are also proud of being able to integrate the rating the user chooses with the different designs our filter offers for the Instagram posts.

What we learned

  • The team learned to appreciate several previously unexplored patches. For instance, the “Vertex Attribute” patch when used alongside “Fragment Stage” provides a surprisingly versatile way of texture manipulation. Likewise, the “Blend” patch can be used to reduce the overall number of materials and objects.
  • The team was also able to discover the hidden limitations of Spark AR. This includes the previously unknown maximum number of patch independent variables, as well as the filter size limit and the importance of compression.

What's next for the project

  • For the next steps of the project, we are going to provide more filter designs. We are also going to improve the generation of the discount voucher presented at the end of the filter.
  • We also plan to present our work to different shops for us to customize the filter according to their needs.

Built With

Share this project: