Umbrella IT

Goal

• To develop a platform for data aggregation and processing.
• The system shall support contract customization and aggregate information around the ‘Client’ entity.
• To create an API providing access to other systems’ data.
• Data shall be collected and processed without delay.

Obstacle

• A huge amount of heterogeneous information located in several sources requires manual processing.
• There is no opportunity to customize contracts.
• There is no aggregated data on all customer’s contracts: only one contract data can be accessed.
• The aggregated data (by customers or by a period) are collected manually.
• The manual data processing does not allow to react in real-time.

Solution

• We unified the data structure and format.
• The user can customize contracts by more than 15 parameters.
• The data are available with minimal latency.
• The aggregated data are automatically collected into a single report.

Goal: To extend dating app functionality.

Obstacle: The existing app architecture allowed the load of 2k users, whereas the business required 100k users.

Solution: We changed the architecture so that the application withstands the required load and asynchronous interaction for a real-time chat. The functionality was extended: the real-time chat and more accurate profile settings were added. The front end was developed in accordance with the new design.

Goal: To develop from scratch an application for creating time-lapse videos from captured clips or photos with the possibility to export them to social networks.

Obstacle: As a video was to be created from several photos the smooth transitions from one frame to the next one were required.

Solution: When overlaying photos, we made each previous photo as a semitransparent layer for the user to align the next photo. Thanks to this, the user manually stabilizes the video and receives a smooth frame transition.

Goal: To design and implement a solution for mobile traffic analysis as an integral part of the client’s team.

Obstacle: It was necessary to create a solution for mobile traffic analysis that had no analogues on the market and ensure its scalability.

Solution: We developed a mobile traffic tracker with instant post-click analysis and an option to set up user redirection to the chosen advertisement. The platform includes a traffic distribution system by geolocation, operating system, mobile device type, mobile operator, as well as the ability to build visualized reports.

Goal: To prepare a landing page for an advertising campaign to be launched by Disney together with SHE Media.

Obstacle: The landing had to be created in a very short period – less than 2 weeks.

Solution: We brought in an IT Reсque Team. We developed the landing page integrated into SHE Media ecosystem. We set up a tag system for the Google DFP platform and created a customized quiz for fans.

Goal: To develop a new paywall and add an archive subscription.

Obstacle: Variety’s previous paywall was developed as part of a monolith.

Solution: 

• We developed the new paywall connected to the core via API and removed the old one.
• A counter tracks a number of the articles read by users and displays a notification upon reaching the limit.
• The unsubscribed users can read only a section of the article.

Goal: To move 600,000 users from BlogHer to SheKnows and update information about their profiles.

Obstacle: It was necessary to merge accounts that doubled each other without information loss. To provide the centralized user management regardless of the platforms where they were registered previously.

Solution: All the users were moved from the old database to the ID Manager, that redirected to the global SheKnows oAuth service. 7,000 user profiles were merged with their duplicate accounts.

Goal: To update the existing data infrastructure of the company after 5 years without an update.

Obstacle: During its existence in the market the project accumulated a lot of data. The audit revealed that 70% of this data was not used. It was nonetheless important to store this data for future use.

Solution: We checked the relevance of the data. Following this, we archived about a petabyte of data to reduce costs and transferred it to another, cheaper service. As a result, the cost of supporting infrastructure and data storage decreased by about 20 times. Data updating and transferring were performed within a month.

Goal: To develop a service allowing to select colored contact lenses online based on a client’s photo.

Obstacle: It is necessary to correctly recognize the iris on the downloaded photos taking into account head movements and position of eyelids and eyelashes.

Solution We analyzed the solutions offered on the market and chose the most appropriate platform. We integrated VisageSDK into the CMS allowing to upload the photo and to virtually select colored contact lenses prior to purchase. 

Goal: To detect whether the franchise partner report has been submitted.

Obstacle: Each franchise uses its own platform to generate sales reports, and they work in different ways. There wasn’t any universal tool to track the reports.

Solution: We developed a tool that analyzes the FTP-server for the data submission from the franchise partner system. Based on the results the tool generates the Missing Sales Reporting Files. The Hamleys administrators have access to the report and can check whether the required data have been submitted or not

Goal: To modify the platform to support high loads and improve performance to make loading faster.

Obstacle: StyleCaster witnessed a drastic growth in the number of users (50 mln users per month). The website had to be optimized to the higher load saving the various site content (images, videos, graphics, audio content, etc.).

 Solution: 

• We performed lossless compression of the content.
• Then we took some load off the server by providing an extra reverse proxy server.
• We partially refactored code to extra accelerate the speed of the website.