According to a World Bank report, major Indian cities are projected to grow faster than anticipated and are going to add another 25 million people to its already dense populace. by 2030.
50% of India’s population is projected to be residing in urban and peri-urban regions by then. thus, Policies like Make-in-India, Ease-of-Doing business which encourage manufacturing and businesses inadvertently nudge industrial units to be located in the peripheral regions of metro cities.
(sometimes) outside the limits of municipal corporations will result in straining the already dilapidated infrastructural resources of these corporations, since extending them to such newer territories will be a challenge in logistics and monetary terms. Planning, under such a constrained environment shall not only focus on the present scenario. but also take into consideration these future projections.
Machine Learning aspects in daily life
One of the most crucial aspects of this new planning would be to ensure the availability of quality, hygienic and reliable water distribution system. though, we look into the various aspects of this mammoth task. how Machine Learning can help address some of them and how it can be synchronised with other government schemes like Smart cities and Jal Jeevan mission to amalgamate them into a comprehensive strategic plan.
Before initialising such a project, a basic understanding of the particular region, its demography, natural and human resources, developmental needs, available infrastructure, economic structure et al would be required to be undertaken, since a one-size fits all strategy will neither be applicable nor be prudent. The project flow could then be summarised as:
Workflow
An initial detailed report of local Lakes, rivers and other water bodies also shall be prepared to record the inventory of water resources in the region. however, along-with already existing water distribution systems, their capacity and distribution network.
Hence, This should be followed by taking stock of existing local demand, classified into residential and industrial categories. so, This initial assessment of demand and supply would aid in better planning and predictions thereof for augmenting the distribution systems.
A regional development plan would be a dynamic document, with cumulative causation. Hence, it will develop continuously over the years. So, Documentation of concrete policy outcomes and their temporal distribution on regional development, would be a challenging task. therefore, This is where Machine Learning would come to our rescue.
Collating data related to progressive set-up of industrial units in an already industrialised region. The region selected must have the following characteristics:
- It must have a wide variety of industries, should not be limited by sector-specific industrialization, so as to ensure data related to various industries and build a better machine learning model.
- Most importantly, it should also have a well documented data corresponding to our needs.
Thus, The industries of the region, their temporal distribution and other specific details like scale of operations will be the feature set for training our Machine learning model. So, Corresponding annual water demand in the same region would be label set for the same model.
A sample table would look like:
| Industry type | Scale | Year | Water demand (annual units) |
| Chemicals | 1.2 | 2012 | 5 |
| Furniture | 0.8 | 2012 | 10 |
| Iron and steel | 1.4 | 2013 | 40 |
| Beverages | 1.6 | 2013 | 60 |
| Chemicals | 1.2 | 2013 | 65 |
| Automobile | 1.1 | 2013 | 80 |
| Paper and pulp | 1.5 | 2014 | 110 |
| Electrical machines | 0.9 | 2014 | 120 |
| Food processing | 1 | 2015 | 130 |
| Plastic manufacturing | 1.2 | 2015 | 140 |
| Rubber | 1.4 | 2015 | 160 |
| Tannery | 0.8 | 2016 | 190 |
| Automobile | 1.4 | 2016 | 215 |
Industrial development plan | Machine Learning
ultimately, building a machine learning model over this training dataset. it can also be implemented on the industrialisation development plan for the region under our consideration, which functions as our test data. The predictions so achieved with this model will guide us in planning for establishing an efficient water distribution network, customised to the needs of the region and its industries, consequently augmenting resource allocation to local municipal bodies to match the demand, leading to a reliable and good quality water availability and thus ensuring streamlined policy implementation.
This capacity planning and efficient resource allocation will provide us with information regarding the available distribution network-capacity levels in real time, which could then be upscaled to exactly match future demands of new industrial units based on this model and hence also guiding us in prospective planning for future industrialisation.
In the last leg of this strategy paper, we discuss the ways and means of aligning this machine learning strategy with existing government schemes like Jal Jeevan Mission & Smart cities.
The two components of Jal Jeevan Mission (JJM) viz.:
- Provision of supplying thus piped water to all rural households by 2024.
- Integrating demand and also supply side water management at local level.
Our ML strategy can become a potent tool in implementing both these goals of JJM, with an added advantage of precise, efficient and real-time resource allocation.
As for the SMART cities mission, a predictive analysis of future water demand will empower municipal bodies in better planning of their monetary and infrastructural resource allocation, while also leveraging these predictions in ensuring a more concrete regional development.
Written By: Sachin Shastri
Reviewed By: Vikas Bhardwaj
If you are Interested In Machine Learning You Can Check Machine Learning Internship Program
Also Check Other Technical And Non Technical Internship Programs
