Big data, Machine Learning and Cloud experts

Data Reply is a Big Data and Analytics consultancy in London, UK. We specialise in two main areas: 

Data Engineering: Desiging and building robust data workflows in technologies such as Google Cloud Platform (Dataproc, Dataflow, BigQuery, Pub/Sub) and leading open source big data frameworks (e.g. Apache Spark, Apache Flink, Apache Kafka, Apache Airflow).

Data Science: Applying machine learning and advanced analytics to solve difficult business questions in areas such as fraud detection, customer insight and lean process management. We have a team of experts with experience in various relevant techniques (e.g. graph analyticsnatural language processing, deep learning) and technologies (e.g. tensorflow, Spark MLlib).

We also offer technical training and advice.

 
$5,000+
 
$100 - $149 / hr
 
10 - 49
 Founded
2015
Show all +
London, United Kingdom
headquarters

Portfolio

GLOBAL DATA PLATFORM FOR VITOL

We designed a global Data Acquisition and Provision Platform for this multinational trader and distributor of energy products.

To read more please see: http://bit.ly/vitol-data-platform

BENCHMARKING TECHNOLOGIES WITH GEOSPATIAL BIG DATA

Features inherent in geospatial data make large scale processing difficult. We helped DSTL to benchmark a range of Big Data technologies that seek to address this challenge.

To read more please see: http://bit.ly/geospatial-big-data

Reviews

Sort by

Data Analytics & Machine Learning for Grocery Store Chain

“Google strongly recommended them as a partner who had experience in the field.”

Quality: 
4.0
Schedule: 
4.5
Cost: 
4.0
Willing to refer: 
4.0
The Project
 
$50,000 to $199,999
 
Nov. 2017 – Jan. 2018
Project summary: 

Using BigQuery, Data Reply compiled three datasets in the cloud to collect information about the relationship between the quality of home deliveries and customer retention. The project was a proof of concept.

The Reviewer
 
10,001+ Employees
 
United Kingdom
Analytics Manager, Food Retail
 
Verified
The Review
Feedback summary: 

Data Reply succeeded in merging the data as desired and helped to identify customers needing higher quality deliveries. They also demonstrated the possibilities that the analysis could bring in the future. The team delivered strong expertise and excellent project management and communications.

A Clutch analyst personally interviewed this client over the phone. Below is an edited transcript.

BACKGROUND

Introduce your business and what you do there.

I’m the analytics manager for a food retail company.

OPPORTUNITY / CHALLENGE

What challenge were you trying to address with Data Reply?

We wanted to analyze multiple datasets to establish how the quality of the home food deliveries was impacting the retention of our e-commerce customers. We needed to combine big analytic data with our order fulfillment system data and our database that showed the history of the customers. This would allow us to see which deliveries were late and how many products weren’t delivered as requested, and thereby assess the quality of the delivery. From the customer database, we could see how many deliveries the customers received. The challenge was to merge three different datasets and to infer the correlation, if there was one, between the quality of delivery and the retention of the customer.

SOLUTION

What was the scope of their involvement?

Their work fell into three broad categories. The first category was account and project management with engagement and fact-finding, and they also acted as liaisons.

The second category was helping us manage the technical details, serving data from our systems into Google’s BigQuery tool. We already had Google Analytics data in there, but we needed the two internal data sources ingested into there as well. We have a technical expert helping us choose which data we needed, the format in which to export them, and the mechanisms for exporting them.

The third category of work was the modeling and machine learning skill set, which took that data and built a risk matrix around it to assess or indicate the level of risk that that customer was perceived to be in based on the three years of history that we uploaded. They built a random forest model that outputted that.

In the end, one output was a technical file that had every customer’s ID and a number indicating if they were a high or low risk. The second output was a presentation to my team and me on what they had done.

What is the team composition?

There were three people. One person led the effort to get the data out of the office and into the BigQuery. There were at least two other people involved in the modeling.

How did you come to work with Data Reply?

Google strongly recommended them as a partner who had experience in the field.

How much have you invested with them?

We spent between $100,000–$200,000.

What is the status of this engagement?

We worked together from November 2017–January 2018.

RESULTS & FEEDBACK

What evidence can you share that demonstrates the impact of the engagement?

It proved what could be done and that it can be useful. Now, practical work is happening to ensure that particular customers get a higher quality of delivery. Hopefully, this will produce real-world results. That’s a relatively small deliverable. This is a proof of concept, which is why we did it. It was designed to prove what could be gained from merging multiple datasets in the cloud and to prove that we could do it. It did both of those things.

How did Data Reply perform from a project management standpoint?

Their documentation was excellent. They gave us weekly reports and we had daily calls. Communication was outstanding. I spoke to a project manager every day to ensure we were on track and aligned. I have no complaints with project management. They came to our office a few times, but it was mainly remote.

What did you find most impressive about them?

The project management and communication were exceptional and clear. They were always available. Their technical expertise is fantastic.

Are there any areas they could improve?

In the early phases, it was a struggle to get a common understanding between their data engineer and data scientists and our subject matter experts in terms of how our system works and what that meant for exporting of the data. I don’t lay that entirely on them; it’s quite complicated and there were difficulties on our side as well. They could work on getting up to speed with a new client’s data more quickly. They should spend as much time as they can early in the project with the people who look after that data.

Do you have any advice for potential customers?

If I was going to run our project again, I would have had them spend the first week or two in our office every day. I’d also block off time for the people in my office to spend time with them every day until everybody understood what that data did. Always do that work upfront and block out as much time as you need.

4.0
Overall Score
  • 4.5 Scheduling
    ON TIME / DEADLINES
  • 4.0 Cost
    Value / within estimates
  • 4.0 Quality
    Service & deliverables
  • 4.0 NPS
    Willing to refer

Analytics Consulting for Fraud Detection Project

“Their commitment to the project was commendable.”

Quality: 
4.5
Schedule: 
4.5
Cost: 
5.0
Willing to refer: 
5.0
The Project
 
$50,000 to $199,999
 
Jun. - Oct. 2017
Project summary: 

As a third-party vendor, Data Reply assisted an internal team with an applied analytics and machine learning project designed to detect suspicious behavior.

The Reviewer
 
10,000+ Employees
 
Ireland
Sector Manager, Inspection Enterprise
 
Verified
The Review
Feedback summary: 

The final report provided deep insights and reinforced strategic decisions. Despite some initial challenges, the Data Reply team were very professional and flexible to the client needs. They acted like a true partner.

A Clutch analyst personally interviewed this client over the phone. Below is an edited transcript.

BACKGROUND

Introduce your business and what you do there.

I’m a business manager for a large inspection, verification, testing and certification organization. We have 10 business lines, one of which is transportation. We have a lot of government contracts around the world that deal with all sorts of inspection and testing of regulated services. In Ireland, we provide inspection systems and support activities for a number of government agencies, including front-office verification for driver licenses processing activities.

OPPORTUNITY / CHALLENGE

What challenge were you trying to address with Data Reply?

We needed analytic consulting for a large government contract, but we didn’t have direct expertise internally.

SOLUTION

What was the scope of their involvement?

Data Reply helped us with an applied analytics and machine learning project to analyze data and identify any potential suspicious behavior. Our client wanted to explore historical data and see if they could detect potential suspicious activity. They also wanted to try out different tools that could identify or confirm possible suspicious activity. 

What is the team composition?

They assigned a data scientist and a data engineer to the project. We also worked with a Partner responsible for project oversight.

How did you come to work with Data Reply?

Data Reply was a referral from one of our team who had successfully worked with them on another contract. Based on the experience they have in transportation in other European countries, we thought they would be a good fit for the project. 

How much have you invested with them?

The project was approximately €48,000 ($57,202 USD).

What is the status of this engagement?

We worked together from June 2017 until October 2017.

RESULTS & FEEDBACK

What evidence can you share that demonstrates the impact of the engagement?

While the project went on longer than originally planned, Data Reply’s input proved to be quite significant. The detailed reports allowed the client to get a deeper understanding of the data. Their output also reinforced and added credibility to decisions the client made on other projects.

How did Data Reply perform from a project management standpoint?

There was a detailed proposal and specifications for the work wi th well-defined timelines, and deliverables distributed across four sprints.


We communicated principally via calls and Skype sessions. I spoke to my contact regularly to see how things were progressing. We had some initial challenges, as it took a while for the teams to knit together and understand the final requirements.

What did you find most impressive about them?

Their team was invested in the project from the outset and in developing a partnership. It wasn’t just about submitting a report; it was also about understanding our client’s goals and managing expectations, which can be challenging when you’re dealing with data. Initially, the client expected results fast, but as the project progressed we worked well together to manage the client expectations. Overall, their commitment to the project was commendable.

Are there any areas they could improve?

There was a learning curve on all sides. For example, we all could have managed expectations better during the early stages of the project. Technology projects can become challenging, especially when they require three-way communication.

Do you have any advice for potential customers?

Goals can change, so you have to be responsive and flexible. I definitely recommend being mindful of that.

4.5
Overall Score
  • 4.5 Scheduling
    ON TIME / DEADLINES
  • 5.0 Cost
    Value / within estimates
    The project went beyond our timeline, but they still charged their original price.
  • 4.5 Quality
    Service & deliverables
  • 5.0 NPS
    Willing to refer