Mobyride
Enabling an e-bike platform to
predict demand
UI/UX Design
Data Engg & Analytics
Mobile App
Mobyride is an Indian dockless bicycle and e-scooter sharing service. It aims to promote sustainable and eco-friendly transportation by providing easy access to bikes and e- scooters through a mobile app.
Industry
Bicycle Retailer
Company size
<100+ Employee
What we did
- Data Engg & Analytics
- Mobile App Development
Objective
Mobyride is a company looking at solving the micro-mobility problem in India. Public transport only gets you to hubs, but Mobyride smart e-bikes and e-scooters gets you to your destination in a cheap and convenient way. The company is trying to expand their markets and make their operations much more efficient with the help of AI. A lot of data has been collected through their connected scooters and their apps. The company looks to use this data to make smarter decisions on improving operational efficiency, expansion strategies and customer experience.
Real-time visibility of their fleet along with statistics such as charge percentage, last charged, battery health and GPS information.
Figuring out maintenance windows of the fleet based on commuter usage and trip feedback.
Improving customer experience by having the vehicle available at the right time and the right place for any customer by predicting commuter demand and frequent routes.
Equip support staff with better data to help customer requests
Challenges
Mobyride is currently collecting a lot of data through their fleet of e-bikes and e-scooters. Also, customer touchpoints such as their apps and website also provide data on customer behaviour. They currently don’t have any way to track the real time data of their fleet. They download the data manually from their devices and batch process this data in their BI tools to gather insights.
This leads to a lot of delay in time from insight to action. They want to be able to cater yo their customers demand in realtime. And this is not possible with the current infrastructure they had started with.
Month to gather data
1
High churn due to unavailability
Solution
Real-time Analytics
Synctactic’s platform supports streaming data as an input source and as an emitter as well. In order to get a real-time view of the fleet’s data the the client followed the below steps
- Added the input source for the streaming data. Here they were using Kafka.
- The Kafka stream was accessed and the individual topics was used to make further transformations using the code engine. The data was stored in a MySQL for further batch processing and the stream data was streamed and transformed further.
- The transformed Kafka stream was then used as an emitter that was connected to a real-time dashboard to view the real-time statistics of the fleet.
Predictive Maintenance
The cold store of the stream was used to understand patterns and train algorithms to predict if a vehicle is due for service or not.
- The MySQL DB was an input source for the fleet data that was fed as training data.
- This data was transformed using the code engine to build out the features for the machine learning model.
- The pre-processed and cleansed data was now fed as an input to an ML model on the platform which provided an output as to when the vehicle would be due for service.
- This was then scheduled to run at some intervals to get suggestions on which vehicles in the fleet are due for service and thus helping the operations plan out their capacity for all the maintenance work.
Demand Forecasting
Predicting demand will help improve ROI on the fleet’s operations. This would also mean that the chances of a customer not having a vehicle available in their vicinity is reduced greatly.
- The customer transactions DB was added as an input source here. The MySQL DB was easily configured and added to the platform.
- The data was then transformed and processed to engineer the features required for the machine learning model.
- The historic data of transactions were then fed to the model as training data. The model provided what times and what vehicles are in demand in what locations.
- The incremental transactions on the DB was read and fed into the model as training data which helped in improving the model as an when new transactions happen on the system.
Customer Data Access
Synctactic makes it easy to access any data in the system by indexing all metadata whenever a new data source is connected.
- Added multiple data sources like support CRM’s, transaction DB, web and mobile analytics to the platform. From here, synctactic automatically extracts and indexes the metadata, hence making it searchable.
- Specific teams and individual members were given access to the data. They then began using the search interface to search all the data in the system.
- Support staff with access to the customer transaction data quickly looked up the customer’s ID and understood their past trips, vehicles used, total revenue generated, etc.
Outcome
The platform hence was used in a variety of ways by the company’s data team to unlock the potential of the data they had been collecting from the fleet and its users. Thanks to the simplicity and ease of Synctactic’s set-up, the team worked on core business problems of the company without worrying about the data infrastructure required to run company- wide data initiatives.
Realtime data availability
Improved team productivity
Recent Work
Launch your product with Arokee
Partner with us for a digital journey that transforms your business ideas into successful, cutting-edge solutions.
