Digging the Business Value Out of Big Data Systems

Nekso Logo

About Nekso


Canadian startup Nekso is a taxi hailing platform that integrates passengers, drivers and established taxi companies into a mobility ecosystem based on the smartphone. The service provides a competitive alternative to Uber for taxi companies and their drivers, and offers better safety and security valued by passengers.





4 Latin American countries

22 cities









Digging the Business Value Out of Big Data Systems


Machine learning, predictive modeling, and other advanced analytics applications help extract the business value out of big data systems – here’s how we helped Nekso build a custom solution from the ground up.

The Challenge


As Nekso’s operations rapidly grew, so did the amount of data generated by its systems. With an ever-increasing supply of data came the challenge of unlocking the information’s value. This issue grew as Nekso pushed to increase taxi-hailing service user growth, which demanded more efficient ways to explore and use more data.


Nekso asked its technology partner, Blanc Labs, to create a big data and analytics solution to reliably collect and analyze large volumes of data.


“We decided to go to Blanc Labs to help us. We could trust their development acumen, data science expertise, and their understanding of our business.” — Azam Mohabbatian, VP Strategy at Nekso


The Strategy


Nekso had four major goals:


1. Gather all their data together in one place: operational, financial, marketing.

2. Consolidate reports and analyses from multiple analytics tools in one place, and create the ability to mix and match reports for a better understanding of their business.

3. Use live reports and dashboards to monitor the business in real time.

4. Use data visualization to comprehend information quickly, identify relationships and patterns, and pinpoint emerging trends.


As Mohabbatian described, “Our goal was to bring all of our information together and then make sense of it – and that is actually a very challenging problem.”


After the discovery phase, the data architecture’s key components were established. Nekso needed a solution that could satisfy a list of business requirements:


  • Capable of ingesting data from multiple sources
  • Designed for scalability and performance
  • Secure, traceable, and auditable
  • Support rapid development and deployment of new applications
  • Allow real-time data analysis with automated feeds to a decision-making system
  • Enable users to explore datasets through visualizations in a dashboard format
  • Offer self-service environments, so that business users can conduct their own data discovery, without having to wait for their IT department.


If the data architecture met these requirements, Nekso could unlock the value in its data.


The Solution


Blanc Labs developed a practical and multifaceted big data architecture that met the business needs. The core of the solution is a data lake hosted on the Amazon Web Services (AWS) cloud platform (Diagram 1).


Data is ingested into this lake from internal taxi operations and from relevant external sources. For example, local event, social media, weather record, and traffic pattern data are used. The flexibility of the data pipelines allows new transformations to be triggered seamlessly on-demand.


Diagram 1: Nekso’s Data Architecture & Advanced Analytics Solution

Diagram 1: Nekso’s Data Architecture & Advanced Analytics Solution


Applications of Big Data Analytics


The data that enters the lake is transformed and analyzed. Advanced predictive analytics forecast trends enabling management decisions and actions. Here are just some of the ways predictive analytics are used:


  • Forecasting and Planning: Temporal patterns in passenger ride requests are identified to forecast ride demand, enabling Nekso’s business development team to better manage and expand their driver base into the best new markets.
  • Operational Efficiency: Ride data is tracked to forecast trends and more efficiently position drivers to meet immediate demand. This allows Nekso teams to more easily identify reasons behind “out of the ordinary” changes in rider behaviour, and take proactive measures to maintain and improve customer satisfaction.
  • Personalized Marketing: Data enables marketing campaigns to be more precisely targeted than ever before. The marketing team can speak more intelligently to specific groups of customers, while testing and tracking their campaign efforts. The ROI of marketing campaigns has increased, expanding the customer base and growing revenue.


Visualizing Business Intelligence Results


The transformed and analyzed data is presented to business users in a customized business intelligence dashboard. The company now has rapid awareness of service performance and KPI monitoring to drive business improvements.


Internal users have access to personalized data and tools that fit their job requirements and expertise.  They can independently explore data, perform ad-hoc queries, and compose their own reports.


“We can now visualize user growth patterns, better understand customer behaviour, and fine tune marketing campaign performance. We’re able to dig significant value from our data.” — Ramón Garcia, CMO of Nekso.


Machine Learning Accurately Predicts Demand, Increases Business


Taxi drivers in the past relied on intuition to find their next passenger. Nekso taxi drivers now have additional tools to locate customers, thanks to machine learning.


Blanc Labs has trained a machine learning system to predict where localized ride demand will be. (Diagram 2). Drivers receive context-aware recommendations on their mobile phones, giving them enough time to reposition their vehicle based on anticipated requests.


The machine learning system uses ridership data, driver location, time of day, weather, social events, and traffic patterns to help Nekso forecast ride demand in a geographic area with 98% accuracy. As a result, passenger wait time has declined, improving the user experience. Drivers and taxi companies both benefit from improved efficiency and higher passenger revenue.


Diagram 2: Real-Time Recommendation System for Driver Location

Diagram 2: Real-Time Recommendation System for Driver Location

In addition, machine learning is used to predict when to send passengers a push notification asking if they will need a taxi. The model can also predict which make and model of car the passenger prefers.


Looking Forward


Blanc Labs helped Nekso aggregate and enrich their data, gain strategic and actionable insights, and improve business results.


“We now have a better understanding of what is happening both inside and outside our business. We know what’s going on at the street level, we’re in tune with our drivers and passengers.” said Garcia.


The results have directly impacted revenue growth for Nekso, and the taxi companies and private drivers using the platform. All users have reported higher satisfaction, particularly passengers.


Interested in working with Blanc Labs to extract value from your data?