Market Research

Robust, scalable data pipelines with AWS infrastructure

Robust, scalable data pipelines with AWS infrastructure

Robust, scalable data pipelines with AWS infrastructure

Processing up to 300 GB of audience and location data every day through scalable AWS pipelines for hourly reporting.

Processing up to 300 GB of audience and location data every day through scalable AWS pipelines for hourly reporting.

BY ARDENT

3 mins read

BUSINESS IMPACT

Real results at scale

  • High accessibility of data

  • Robust pipeline infrastructure

  • Speedy processing of large volumes of data

  • Expert technology recommendations

The Challenge

Our client were looking to collate large amount data from various sources with high data accessibility and accuracy to provide a speedy turnaround time for their end customer.

The Solution

Well-established market research company

Our client are a well-established market research company offering a number of services including market research insights on tv audience data. The company has been providing data to end customers to help them make well-informed decisions on marketing, business strategy and approaching their target audience. With almost 35 years of experience, they look to expand and grow their services and we are proud to help them on their journey to achieve that.

Collecting, processing and enriching a variety of data

We have worked with our client on a number of data projects, so when they needed robust, scalable pipelines built to deal with large volumes of data, Ardent was the data engineering provider of choice. Our client wanted to ingest data from several sources, collate it and enrich it for their data science teams to generate deeper insights that inform smarter decisions for their end customers.

Robust, scalable pipelines

Our engineers with their rich experience and expertise in delivering data engineering excellence for over a decade built robust, scalable data pipelines to deal with the large volumes of data being ingested.

We built the data pipeline infrastructure to collate data from three main sources including TV data from a leading provider of automatic content recognition technologies and cross-screen metrics, mobile data from a global leader in mobile location data, POI data and corresponding compliance services as well as pixel data.

Each source of data provides the data in various formats, which was cleansed, processed, validated and enriched by joining tables based on set queries with intelligent algorithms. This provides the client with a fuller picture of their data.

One of the key technical challenges that our team overcame was dealing with large volumes of mobile data which varied between 100 GB to 300 GB per day. This data needed to be processed swiftly, therefore our teams were able to recommend and employ EMR (Elastic MapReduce) to significantly reduce the processing time so our client could access the data quickly.

High data accessibility and data accuracy

Our client can access hourly reports based on their data, which is processed quickly and efficiently. This allows their data science teams to provide quick turnaround to their end client on data reports and insights, empowering them to provide excellent service to their clients. We continue to work with our client on managing their data pipelines with our operational monitoring and support service.

The Results

The AWS pipeline processes and enriches high-volume multi-source data and gives the client hourly reporting, helping its data-science teams provide faster insight to end customers.

Looking to achieve similar results?

Talk to our team about how we can help.

Get in touch

INDUSTRY

Market Research

SOLUTION

AWS S3, Athena, CloudWatch, CloudFormation, CloudFront, Managed Airflow, EMR, AWS Batch, SNS, EC2, Bitbucket, Airflow

KEY FACTS

100–300 GB

Mobile data processed per day

Hourly

Reporting available from the processed data

3

Main data-source groups ingested

Ready to make your data work harder?

We build and manage the data foundations that remove the quality, consistency and reliability issues holding back AI, automation and better decisions.