Navigating India – The Journey of Ola Maps

Mapping is a core pillar for mobility. They’re not just about getting from point A to point B; they’re about understanding the intricate web of streets, traffic patterns, and urban landscapes that define our cities. Humankind has always relied heavily on advancements in mapping technology, to progress in life.

In 2021, as the world started to move again in a post Covid era, we had plans to build first-to-world features on our customer and partner apps in Ola Mobility that would transform the user experience, but we quickly hit a roadblock. Our reliance on western mapping providers, for whom India wasn’t a priority, severely limited our ability to implement these features. We realized that to truly serve our users and push the boundaries of mobility, we needed a true alternative to global mapping giants—one that was not only better suited for the Indian market but also more responsive to our unique needs. This realization sparked the inception of Ola Maps.

But why build our own maps when there are existing solutions? 

The answer lies in our vision for the future of commerce and mobility. We’re not just building maps for today: we’re creating the cartographic infrastructure for tomorrow’s mobility innovations. From autonomous vehicles to flying taxis to drones, the future of mobility demands maps that are more detailed, more dynamic, and more intelligent than ever before. We embarked on this ambitious journey three years ago, recognizing that to truly revolutionize mobility, we needed to own and innovate on every aspect of the maps experience that power our platform.

This post kicks off our 4-part series on Ola Maps. Here is what’s coming: 

Part 1: The Foundation of Ola Maps  (You are here)

Part 2: The Data Behind the Maps

Part 3: The AI Revolution in Mapping

Part 4: The Future of Maps

In the following sections of this blog, we’ll dive deeper into the technology behind Ola Maps, the challenges we’ve overcome, and how this initiative is set to transform not just Ola, but the entire cartographic landscape. Stay tuned as we map out the future.

Section 1: Challenges & Opportunities while Building Maps for India

Existing mapping providers do not fully address unique challenges related to delivering a seamless experience for Indian users and this creates special opportunities for Ola and Indian developers. 


  1. Incomplete Mapping Coverage: Many streets and rural areas are not mapped or poorly mapped.
  2. Inconsistent and Varying Street Names: Street names often have variations and inconsistencies, leading to confusion.
  3. Frequent Changes in Road Networks: New streets are created, and existing ones are closed frequently, leading to outdated maps.
  4. Traffic and Road Condition Variability: Traffic patterns are highly inconsistent, and road conditions vary significantly due to potholes and non-standard streets.
  5. Non-Standardized Streets: Many streets do not conform to standard measurements and layouts.
  6. Potholes and Road Quality Issues: Potholes and road quality can significantly affect travel time and safety.


  1. Enhanced User Experience: Providing accurate and up-to-date maps can significantly enhance navigation and travel planning, improving user satisfaction.
  2. Localized Features: Offering features tailored to local needs, such as multi-language support, local business listings, and culturally relevant landmarks.
  3. Smart Navigation Systems: Developing smart navigation systems that can suggest optimal routes based on real-time data, considering traffic, road conditions, and user preferences.
  4. Integration with Smart City Initiatives: Collaborating with smart city projects to integrate mapping data with other urban infrastructure, enhancing overall city management and services.
  5. Augmented Reality Navigation: Utilizing AR to provide intuitive and immersive navigation experiences, especially in complex urban environments.
  6. Real-time Predictions: Using state of the art time series models to foresee traffic congestion, road closures, and other disruptions, offering proactive routing suggestions.
  7. Community Engagement: Engaging with local communities to crowdsource map updates and feedback, ensuring the map remains relevant and accurate.
  8. Eco-friendly Routing: Implementing eco-friendly routing options that minimize fuel consumption and emissions, appealing to environmentally conscious users.
  9. Real-time Data Updates: Implementing real-time data updates through IoT devices, GPS tracking, and user-reported changes can keep maps current.
  10. Local data residency: Ensure that critical data like location intelligence never leaves Indian soil

Addressing these unique challenges is not easy. In order to address them and tap the opportunities, we are leveraging three core resources: the power of AI, the power of open source and the power of the vast Indian talent ecosystem. This approach not only allows us to be more efficient but also ensures that our maps are built for delivering contextually relevant, accurate, safer and enhanced customer experience.

Section 2: Ola Maps Overview – Core Foundations

Fig 1: Overview of Ola Map ecosystem where we have multiple data sources being used to build map data layers exposed via Services & SDKs

Ola Maps represents a sophisticated and integrated approach to digital mapping, combining multiple data sources, advanced AI algorithms, and user-friendly SDKs to deliver comprehensive navigation services. While the actual architecture is considerably more complex, we’ve distilled it down to its essential components to provide a clear, high-level understanding of the system’s structure: Data, Places, Tiles, Routing and SDKs.


For data sources, Ola Maps is built to utilize the most diverse set of data and send updates in near real time to ensure the most accurate mapping data as possible. Our AI first data systems utilize real time data from millions of vehicles using Ola Maps, fleet of Ola S1’s equipped with 360 cameras, open-source government data repositories, OpenStreetMap, partnerships and proprietary sources to build essential map features such as roads, points of interest, street furniture, building geometry and traffic signals. This highly comprehensive data system helps us with a robust mechanism for map expansion and map maintenance while allowing us to contribute to the open source community. In the past one year alone, we have contributed a total of 5.43 million edits to Open Street Maps.


Ola Maps has adopted an AI-first approach to revolutionize our places system, responding to the generational shift in search and location-based services. We’ve integrated advanced machine learning and natural language processing to create a dynamic, context-aware places database. This system adapts in real-time to user needs and query intents, moving beyond static POI data to offer intelligent, predictive place information in multiple Indian languages.

Rendering & Tiles 

Ola Map’s tiling system represents a strategic fusion of reliability and innovation. We’ve built upon an open-source rendering stack, leveraging its stability and community support for our base tiling needs. This gives us the strategic space to invest in cutting-edge NERF (Neural Radiance Fields) technology stack, pushing the boundaries of 3D scene reconstruction and view synthesis. This hybrid approach allows us to deliver consistent performance while gradually introducing photorealistic, dynamically rendered environments. 


Ola Maps’ routing capabilities stand out due to their integration of preference-based routing, which tailors routes to user and vehicle preferences, and popularity-based routing, which suggests commonly used paths. Real-time and historical speed predictions provide accurate ETAs, while obstruction and closure management ensure routes are free from disruptions.

2.1 Ola Maps Data Ecosystem

This section gives a glimpse into our data ecosystem where we have different sources fed into our pipelines with AI agents augmented by a small number of highly skilled human data editors focusing on coding map features and updating our data layers.

Fig 2: Overview of Ola Maps Data Ecosystem

Open Source

Use Case


Data layers (Road Names, POI, Administrative areas, Traffic Signs, Traffic Signals, Buildings)

Data Sources

Our data ecosystem is built on the principle of crowd-sourced geo-mapping information leveraging Ola Maps massive user base, Open street maps, open source Government data repositories and other proprietary sources. Our system harnesses the wealth of location information generated by anonymised GPS pings & curated using AI models. The sensor data generated from 2-wheeler, 3-wheeler & 4-wheeler helps in identifying missing roads, real time traffic congestion & road restrictions and many such data points.

Fig 3: Map of Mumbai showing data pings from Ola Maps ecosystem in 1 min

Fig 4: Street Imagery being captured by our data operations fleet. Using Ola’s AI models, geo data (POIs, traffic signs, turns and more) are retrieved from this feed.

Data Platform

There is a highly sophisticated data platform in place which ingests more than 5 million messages per second from a multitude of sensors and telemetry sources. The data is centrally ingested, normalized, anonymized and then put in our Lakehouse hosting petabytes of data. Data streams coming from various sources are divided into further declarative pipelines to collect relevant data streams for training AI models, Analytics, and Data ops for Maps. The final output is put into Maps databases for Tiles, Places and Routes systems.

Data Editor Platform

The Data Editor Platform facilitates the integration, enhancement and continuous updating of mapping data. AI agents augmented by a small number of highly skilled human data editors focusing on coding map features and updating our data layers. 

We will dive deeper into our Data Ecosystem in Part 2 of this blog series “The Data Behind the Maps”.

2.2 Ola Maps Places Platform

Our Places platform built on our POI repository supports the core functionalities of Autocomplete, Forward Geocoding, Reverse Geocoding and Nearby Search.

Fig 5: Overview of Ola Maps Places Platform

Autocomplete: The Autocomplete API dynamically enhances search functionality within various applications by identifying entities in user queries and suggesting relevant completions in real-time, significantly improving the user experience

Forward Geocoding: Geocoding API is a service that translates human-readable place names and addresses seamlessly into geographic coordinates, such as latitude and longitude.

Reverse Geocoding: Reverse Geocoding API is a robust tool crafted to effortlessly translate geographic coordinates into human-readable place names, spanning from precise street addresses to broader geographical areas

Nearby Search: Nearby Search API  is a service that lets users find locations within a specified area.

<Need to add content around the systems from the figure>

Open Source usage in Places Platform

Open Source

Use Case


Autocomplete API Structure

AI enablement of Places Platform

The Places platform leverages a host of custom built advanced AI models to improve user experience.

POI Popularity and Classification

This model ranks and classifies Points of Interest (POIs) based on various factors like user interactions,  frequency of visits, and user feedback. The classification helps in providing more relevant search results and recommendations to users.

Search Relevancy Ranking

This model ensures that the most relevant POIs appear at the top of the search results by using state of the art mBERT (multilingual encoder) based Named Entity Recognition (NER), Learning to Rank (LTR) models and natural language processing (NLP) techniques. The model takes into account factors like query context, user preferences, and historical search data to rank the results effectively. 

Contextual Search Identification

This model understands and processes context-based search queries to provide accurate results. This is done by syntactic and semantic search models like Bi-LSTM-CRF. It also leverages NLP and contextual analysis to comprehend the intent behind user searches. This model uses contextual cues and user behavior patterns to fine-tune search results.

We will dive deeper into how AI has enabled us to build a world class mapping platform at a fraction of time & cost  in Part 3 of this blog series “The AI Revolution in Mapping”.

2.3 Ola Maps Routing Platform

Our Ola Maps Routing Platform is designed to deliver precise and efficient navigation services through a sophisticated data processing and routing system. The platform integrates data from various sources, including real-time signals, to generate accurate direction’s and ETA’s.

Fig 6: Overview of Ola Maps Routing Platform

Directions: Provides turn-by-turn navigation instructions to guide users from their starting point to their destination.

ETA: Calculates the Estimated Time of Arrival based on current and historical traffic data to provide accurate travel times.

Snap to Road: Aligns user-reported positions to the nearest road, correcting any GPS inaccuracies to reflect real-world driving paths.

Map Matching: Ensures that GPS traces are accurately aligned with the road network by matching them to the correct paths on the map.

Multi Stop Routing: Enables planning and navigating routes with multiple destinations, optimizing the order of stops for efficiency.

<Need to add content around the systems from the diagram above>

Open Source usage in  Routing Platform

Open Source Models

Use Case


Routing (Custom MLD algorithm)

AI enablement of Routing

We have built models to enhance our Routing service offering while improving user experience.

Vehicle Pings Analysis: Uses real time data streams from multiple sources to predict road segment speeds, identify roads, traffic conditions and potential disruptions. It helps in dynamically adjusting routes based on current road conditions and traffic flow.

Road Closure Model: Identifies temporary and permanent road closures to enhance routing accuracy by utilizing real time data from government sources, social media and vehicle pings to detect road closures. The model continuously updates the routing engine to avoid closed roads and suggest alternative routes.

Speed Prediction: Leverages neural network based time series models using historical data to estimate travel times more accurately where realtime coverage is less.  

Road Popularity Model: Aggregates data from various sources, including vehicle pings, user preferences and user feedback, road classifications to assess the popularity of different roads. Popular routes are prioritized for routing suggestions, ensuring that commonly used and trusted paths are recommended to users.

2.4 Ola Maps Tiles Platform

<Need to write one line description here like in the other sections>

Fig 7: Overview of Ola Maps Tiles Service.

Tile Generator

The Tile Generator is responsible for processing the ingested data to create map tiles. It leverages the configurations that define the visual appearance, such as colors, line styles, and icons as well the settings that determine how the tiles are generated and displayed, ensuring consistency and usability across different platforms and devices.


Tileserver creates a map instance on the server-side (leverages the Open Source Maplibre library) with local styles, fonts, sprites and tilesets. This setup is used for rendering the base map of static map tiles. A transparent image with markers, polylines and watermark is overlaid on the base map image to produce the final map output.

Open Source usage in Tiles Platform

Open Source

Use Case


(Includes Mapbox & Map-tiler libraries)

Tile rendering, Marker rendering, Feature and style rendering, real time route rendering

Section 3:. Impact on Customer Experience

Ola Maps has been going live across the Ola ecosystem and our users are loving it.

“I love the real-time traffic updates on Ola Maps. They are more accurate than other Maps products that I was using, which saves me a lot of time.”

“Real-time traffic updates on Ola Maps are a game-changer. I can avoid congested areas effortlessly.”

“I find Ola Maps to be more reliable in providing real-time traffic updates compared to other apps.”

“I’ve found a few locations on Ola Maps that were missing on other Maps. It’s great to see such accuracy!”

“Ola Maps has really impressed me with its precision. The navigation is smooth and the directions are spot on.”

“I’ve noticed that Ola Maps has captured several local spots better than other Maps products. Kudos to the team!”

“The ability to transfer locations from my phone to my scooter with Ola Maps is fantastic. It’s a big time-saver.”

“This is how you share a location from WhatsApp or any other app to your scooter’s dash through the Ola companion app.I envision myself utilizing Ola Maps for navigation much more often than Google Maps, thanks to this convenient feature.”
“The user interface of Ola Maps is so clean and easy to use. It makes navigation a breeze.”

“I appreciate how Ola Maps integrates seamlessly with my Ola scooter. It’s a perfect match!”

“The feature to search the location on my phone and transfer it to my Ola scooter is incredibly convenient.”

“So far, my experience with Ola Maps has been nothing but positive. It’s intuitive and reliable.”

“Ola Maps is very user-friendly. Even my parents find it easy to use, which says a lot!”

“The integration of Ola Maps with Ola scooters is genius. It makes the whole travel experience seamless.”

“I just used Ola maps in my S1 Pro and it’s working very well in Pune… UI shows so much details… I would have missed the turn on Google maps but not on Ola… Seeing signals in the map UI for the first time…”
“The turn-by-turn navigation on Ola Maps is excellent. It keeps me on track without any confusion.”

“I’ve had a great experience with Ola Maps so far. It’s accurate and gets me to my destination without any hassle.”

“The detailed mapping and accurate routes on Ola Maps have made my trips so much smoother.”

3.1 Ola Maps in Ola Electric Ecosystem

Every day tens of thousands of customers use Ola Maps on their scooters and the Ola Electric app for a guided commute to places in their cities. The platform serves over a million search queries everyday for Ola Electric users. Use of on-device navigation has gone up 4x since the launch of Ola Maps in MoveOS 4 and in a recent survey, it was voted as the top feature in MoveOS4.

3.2 Ola Maps in Ola Consumer ecosystem

Ola Maps forms the core of our Ola Consumer platform: it forms the base layer of our apps, helps match the rider – driver pair, provides key inputs to our price prediction algorithms, powers all the pick up and drop off experiences and the core navigation experience for all our driver & delivery partners. Our enhanced places platform now outperforms our previous map provider both in terms of accuracy (due to superior user context) and latency. Users love the detailed information about their cab’s route to their location with detailed traffic states and traffic lights powered by the Ola Maps tile platform. The routing engine has significantly improved our pricing prediction and allocation algorithms leading to 2% lesser pricing deviations, 9% lesser cancellations and 6% accurate pickups.

3.3 Comparison of key performance metrics

The Ola Maps platform today outperforms our previous map providers on key performance metrics for the Ola users.


Fig 8: Chart 1: Location & Search accuracy between Ola Maps, Map Provider 1 and Map Provider 2

Fig 11: Chart 2: Search accuracy and ETA accuracy for 2-wheelers between Ola Maps, Map Provider 1 and Map Provider2

3.4 Impact on external map providers

Over multiple launches, AB experiments and product iterations, the Ola Maps platform has outperformed our external map providers across all services, geographies and form factors. With every release, our reliance on external map providers has come down. And in the month of July, it is going to be ZERO! This has allowed us to invest our resources into building significantly better experiences for our customers and to imagine the future of mapping!

Fig 14: Screenshot of external map provider cost dashboard

We hope that you enjoyed reading about our journey of building Ola Maps and the core foundations of the system. Stay tuned for our deep dive into the world of Ola Maps in our upcoming installments where we’ll explore the intricate data foundations, cutting-edge AI applications, and our vision for the future of mapping technology. Don’t miss these insights into how we’re reshaping the landscape of location-based services. All Ola Maps APIs are available for everyone. Visit and start building your applications on APIs that are tailored to enhance location-based services with precision and efficiency for the Indian market.