This post was originally published by Debolina Biswas at Analytics India Magazine
Founded in 2010, Cranberry Analytics is a tech-enabled company operating in the water management space. Headquartered in Pune, the company was co-founded by Shishir Thakur, Amit Deshmukh and Onkar Gauridhar.
Cranberry Analytics was founded with the primary objective of measuring water efficiently. With the use of analytics, ML and IoT, the company augments government and large water infra companies with a tech stack to enable them to curb water wastage and plug revenue leakages.
In a conversation with Analytics India Magazine, Co-founder and CTO Shishir explained the technology behind Cranberry Analytics’ management system Recon, breaking down how AI and ML can facilitate the management of natural resources and what the future for AI in water management looks like.
Edited excerpts from the conversation:
AIM: What is your flagship product/service?
Shishir Thakur: Our flagship product is our water billing, budgeting and management system – Recon, deployed in PCMC (Pimpri Chinchwad Municipal Corporation) since 2012.
Recon integrates with all data sources related to:
- Water consumption
- Water distribution
- Demand management systems (billing)
- Revenue management systems (manual and digital)
- Customer experience and dispute resolution
It also provides a platform and dashboards to all the concerned personnel of the water department, citizens, and support providers.
AIM: How does Recon work?
Shishir Thakur: Recon has built-in capabilities to detect anomalies related to water usage and leakages, revenue leaks, automatic dispute resolution, tracking defaulters, predicting and forecasting water demand, tariff simulation, integrations with smart water meters, sensor management, accounts department and several other auxiliary functions. Using data visualisation in reports, charts and tables, Recon makes it very easy for top management to form policies and implement relevant changes that have the most impact.
AIM: Please explain the tech stack used by the team.
Shishir Thakur: For the Recon frontend, we use Angular.js, Charts.js, Tableau and Metabase. We primarily use node.js with some specialised algorithms in Golang and Python for anomaly detection and regression through past data to find patterns and forecasts for its backend.
- Databases include MySQL, MongoDB and Realm-DB.
- Mobile: Android (with Realm DB) integrates with onspot thermal bill printer and portable PoS machine for revenue collection.
- DevOps: Docker, Ansible, Jenkins, Nagios, Bitbucket.
AIM: How does Cranberry Analytics leverage AI and ML capabilities?
Shishir Thakur: We analyse all incoming data to find consumption related anomalies (similar to how a credit card fraud detection system works), and then alert the respective personnel. This has helped us identify many leakages, unauthorised consumption, geographic segmentation of faults based on geotags, and human errors.
Additionally, we leverage AI and ML to predict area-wise water demand; assist water distribution staff in achieving equitable water distribution with intelligence gathered from several multivariate analyses; for automatic dispute resolution and bill rectification; for payment-related inconsistencies; and for the coordination of field operations.
AIM: Who are your clients? Share a few real-life cases where your product has been used.
Shishir Thakur: We have worked with clients like the World Bank, Asian Development Bank, Karnataka Urban Infrastructure Development Corporation, Ministry of Urban Development, Ministry of Housing and Urban Affairs, and Suez International.
Case Study: PMAY (Pradhan Mantri Awas Yojna)
- For an initiative by the Ministry of Urban Development, we have been augmenting them with technologies, data analysis and software infrastructure to enable them to identify the correct approach towards the slum dwellers, strategies and software for more effective data collection and survey exercises, and identifying the truthful beneficiaries for the program and eliminating the spurious ones.
- Service-level benchmarking: An initiative by MoUD to identify, analyse and improve the services being provided to citizens from government bodies. We helped them in large scale data collection, statistical analysis, identifying anomalies, building models, and defining policies based on that data.
- PCMC water billing management: We helped PCMC, a municipal corporation with a 25 lakh+ population, in reducing their unauthorised water consumption, improving revenues, reducing customer complaints, enabling equitable water distribution and reduction in water consumption to save 31000 million litres of water per year. Revenues have reached from Rs 22 crore per annum to almost Rs 43 crore already, with a 95 per cent reduction in customer complaints and almost nil non-metered connections and historical disputes.
AIM: How can technology further help in water management?
Shishir Thakur: The biggest impetus can come from smart metering of cities, implementation of IoT devices and sensors in water distribution, and rapid growth of smart tech in agricultural water consumption.
The biggest returns on investment can be seen by:
- Implementing technologies like drip irrigation, sprinklers, and air-drop irrigation in farming — the biggest area of ineffective water utilisation today.
- Implementing better systems to detect leakages in urban pipelines and fixing them as soon as possible. We lose more than 60 per cent of treated water due to underground seepages.
- Mandatory implementation of rainwater harvesting in urban areas and effectively monitoring its efficacy using tracking technologies.
- Mandatory waste-water recycling, especially in large apartments, industries and urban centres, and greywater usage instead of freshwater, wherever possible.
- Implementing and experimenting with low power, low cost, novel methods of seawater desalination.
- Use of ML for preventive hydraulic modelling of our current dilapidated storm drain infrastructure for better rainwater catchment to reduce floods and recharge water bodies.
- A strict ban on letting effluent and chemical waste enter river bodies, that’s a huge drain on resources and can be easily monitored centrally and remotely using chemical sensors and IoT devices.
AIM: What does the future of resource management using technology look like?
Shishir Thakur: Historically, all struggles related to resource management and resource scarcity have originated from slow movement of information, deliberate and artificial impedance imposed by governments or large businesses, inability in simulating complex future scenarios due to a multitude of variables, opaque information channels from ruling regimes, lack of resource visibility and operational efficiency.
The more connected we become, both in terms of information transfer and the stacking effect of innovations on top of each other, we keep becoming free of geographic and political constraints. This kind of global collaboration is unprecedented and unpredictable. This pattern will automatically fuel itself to a level where natural market forces will instantly eliminate scarcity of any kind in any area of the world.
Just like it was difficult 15 years ago to envisage the kind of product and food deliveries that we now take for granted, it may seem difficult to imagine this in 2021, but with the use of AI augmentation and other tools built on the ever-improving IT backbone, the natural market economies will prevail and resource management issues of today will become a trivial problem to solve. We will surely face novel challenges of different kinds, but speculation is futile, given the extremely rapid pace of innovation.
AIM: What does the road ahead for Cranberry Analytics look like?
Shishir Thakur: The solutions that we have developed in PCMC are already quite comprehensive. We have plans to expand these and create an Integrated Water Management System (including water billing, water budgeting, water analytics and the instrumentation aspect of it) using sensors and industrial IoT devices to monitor water’s lifecycle end-to-end, right from the source till it ends up in the waste-water treatment plant. We are in talks with a few state-level entities for a larger implementation of these solutions. Because, even at such a small level, due to our efforts here, PCMC is now saving potentially ~31000 million litres of water annually, it becomes imperative for us to let the rest of the country benefit from the same processes.
Currently, we are also looking for investors who are aligned with our mission and can provide us with the wherewithal of rapid expansion into not just the rest of India but also in water-scarce places like the Middle East and Africa. We have a pilot project being implemented in Kenya and are talking with a few Dubai entities.
To understand if AI can actually help in resolving India’s water crisis, check this article.
This post was originally published by Debolina Biswas at Analytics India Magazine