Smart Aid for Self-help (SAS)

The Self Help Groups(SHGs) movement in India provided opportunities to economically empower and socially engage women better. But not all SHGs succeed in terms of sustainability and profitability. Lack of cross group learning, digital & financial literacy, process delays, and unhealthy business practices leads to SHG failures. Smart Aid for Self-help(SAS) aims to pilot technology-enabled SHG administration to support a collaborative mode of information management, knowledge generation  and dissemination thereby promoting the SHGs towards success. The objective of the project is to digitize all SHG activities with mobile applications and provide enhanced information-based services to the stakeholders supporting policy formulation and project implementation. The decisions making can be prior informed by the effective decisions undertaken by other successful groups. The system can timely advocate the SHGs with potential connections with suitable business stakeholders. Bulk amount of data generated through the digitization would set stage for knowledge creation and policy formulation.

Women in rural communities of India often have least access to monetary resources making them more vulnerable at the face of adversities. As of March 2019, India has around 44.61 lac women SHGs and in Kerala, Kudumbasree network (one of the country’s earliest community platforms of women) alone has 4.4 million members. AmritaSREE is another program already successfully implemented in rural areas throughout India through several projects by Mata Amritanandamayi Math (“MAM”) and affiliated to Amrita Vishwa Vidyapeetham (AVV). It provides basic microfinance training programs to 250,000 women primarily in South India as well as training rural women across the country in vocational and life skills thereby providing capacity building and employment. From our experience we found that the success of SHGs often depends on factors like social cohesiveness, participatory decision making, equal access to the benefits & information, consistent tracking of the transactions etc. Constrained by many challenges, certain communities find it challenging to sustain the motivation and participation from the women. Even if a group of women manage to kick off with an SHG, their lack of exposure in establishing proper linkages with the stakeholders and running successful marketing, leads to eventual failure of the SHGs.

We propose a cloud-based software platform comprising mobile applications, web applications, and cloud services to digitize SHG activities and enhance its impact by propelling the SHGs towards sustainability. The main features of the solution are as follows.

1) Digitization of the SHG activities by developing and deploying a mobile application that supports the digital recording of the daily transactions, interactions and decision making. To manage group profiles, maintain digital notes for SHG meetings, guide loan applications, and business procurement services so that the SHGs can effectively scale with minimized chances of failure.  

2) Real time tracking and monitoring of SHG activities through a web platform with customized access to its stakeholders. It provides timely notifications of the group’s activities, performance assessment reports for the SHGs, cross group communication platform and visibility to the potential stakeholders.

3) Data analytics and services to identify the best practices that better sustain an SHG: This comprises modeling the SHGs behaviors, identification and modeling of best practices, and  timely warning for potential threats.

Technology to-be-used in the project include web and mobile applications, cloud storage and implementation of data mining techniques. The data science techniques would primarily include identification of behavioral patterns and optimal recommendations.

The target group of women belongs to the rural communities of India. Despite the multitude of various roles and responsibilities undertaken by them, most of them  are not financially and socially independent decision makers. According to the recent reports by NABARD (National Bank for Agriculture and Rural Development) in March 2019, India has over 44.66 lakh SHG groups across the country. Mata Amritanandamayi Math, has been implementing SHGs across Kerala through its AmritaSREE (Amrita Self Resilient Education and Employment) project. Currently it benefits 2,50,000 women in Kerala, India and has started to scale across other states of India. Our researchers have been conducting structured qualitative and quantitative research to understand the general challenges faced by these SHGs. Data quality, equal 24x7 access to information, process delays, resource constraints and beneficial business connections were the major challenges reported. The features of the software solution have been designed to address the challenges of conducting activities transparently and provide equal access to information. The advantages of digitizing is that it makes everything available for SHG members on a button click, one can conduct a transactions with limited codependency, avoid local travels, readily access the information 24 * 7, and suggested business transactions.

SHGs sustains mainly on factors - Active participation, proper management and healthy linkages with the stakeholders. Digitilization of the SHG activities would enable real-time alerts and notifications to its members that would hopefully motivate them by improved transparency and better access to information. What worked for an SHG in the southern part of India may not work well for the northern. So the data that comes through digitilization can be consumed by data science or AI modules to gain context specific insights on best practices or links to sustain the SHGs over time.

Our innovation is on the digitization of the SHG operations to enable the application of data mining techniques to identify and recommend best monetary and business practices to sustain the SHGs. The intervention will deploy a customized mobile app for the SHG members, that is co-designed with the target population to ensure usability (i.e. local language, simple interface, etc.).  The app addresses many of the challenges currently faced by SHG federations, which lack quality management, financial analysis, and sustainability (Parikh, 2006; Charumathi, 2017, APMAS, 2007, Huma, 2017).  It will support SHG group formation, performance & efficiency, as well as real-time tracking, systematic data management and communication, which has tremendous potential in supporting knowledge sharing with key stakeholders at the government policy, and project implementation levels especially relevant to the COVID-19 scenario. Often SHGs also struggle due to the lack of knowledge about the best financial and business practices that often varies based on the behavioral models and local markets. The SAS application will be deployed among the AmritaSREE SHGs and the digitization data will be leveraged to implement data science techniques to identify the specific patterns in business transactions or monetary transactions that have sustained the SHG in its long term goals. As part of this research would be conducted on SHG behavioral modeling, identification and modeling of best practices, and generating timely warning for potential threats. These features would be then integrated with the app to further enhance it by enabling collaborative information sharing and knowledge dissemination.

1) Cloud based data services would be used for data storage and analytics. We plan to implement our data model with a firebase that supports No-SQL. 

2) Android based mobile applications were designed, developed and implemented. We go for android because in India its the most economic choice for women and it supports tablets too. Unity3D has been used to design and develop the app with C# as the back end.

3) A Python based web portal needs to be developed to serve knowledge dissemination among stakeholders.

4) Data science techniques would be deployed to identify best practice patterns in terms of monetary transactions and SHG activities. Based on these patterns and context relevant information the system can make recommendations on suitable best practices. Data mining techniques like  associative rules and collaborative filtering can be used to model the SHGs behaviors and generate appropriate recommendations.

The mobile application of the SAS application suite is developed for the android platform. The fact that 401.74 millions of Indians are using Android based smartphones, makes the application more available to the common man. The architecture of Cloud based mobile and/or web applications have been quite successfully deployed by many solution providers and has been the core medium of digitalization. Moo Farm is one such solution that serves 100+ villages in India to help the marginalized dairy farmers enter the 21st century farming by using ICT technologies to create sustainable rural livelihoods. The digitalization achieved through their dairy farm management application and leveraging on data analytics and big data to gain impact on productivity and income generation. The farmers track the entire cattle life-cycle, milk production and business to receive real time automated support based on data analytics and also from human experts.

The first set of activities include developing and deploying a mobile app among the SHG women which will evaluate the performance of 75 SHGS, and lead to the short-term outcome of improved SHG management. The multi purpose SAS application suite will support the overall monitoring and control plan that will be accessible to all stakeholders, including SHG members themselves. Receiving a direct assessment of SHG performance should allow the SHG members to improve their practices on their own. App-based information will be supplemented with in-person monitoring at regular intervals by the project staff. This strategy for monitoring, evaluation, and regular communication is an effective way to overcome the challenges of distance and scale, particularly critical in the COVID-19 context.  The second objective of the proposal is to provide equal access to information and effective knowledge dissemination. We can describe the administration hierarchy as follows. An SHG is managed by a women facilitator, more that one SHG group are grouped as a cluster unit governed by cluster officers, multiple clusters overseen by zonal officers and headquarter staff. Every member of the hierarchy should have access to information and knowledge. A web dashboard would be developed to provide customised access to each level of users aligned to the hierarchy. Apart from theis the system should also serve the market partners, and other linkages to improve the visibility of the SHG efforts. This would set the platform for financial and market linkages, and digital communications.  The third objective is to provide cross learning from the experience of other programs and actors within the ecosystem. Stories of success and failure experienced by SHGs can be used to inform the other groups about the good practices and bad practices thereby improving the efficiency of the system.

Currently AmritaSREE (Self Reliance Education And Employment) serves 2,50, 000 women in the domain of microfinancing by the parent organisation, Mata Amritanandamayi Math. These women groups are primarily located in South India.  They have initially reported to have been facing several challenges in terms of accounting, choice of beneficiaries, loan interest negotiation and managing delinquency, and have requested our support on financial practices and business training. As our effort to pilot the SAS application suite, we are targeting to deploy it in the nearby districts of Kollam and Alappuzha spanning around 75 SHGs that is approximately 750 women. In the next two years we plan to scale up the deployment across all the district of kerala to reach out to 2,50,000 women across Kerala. 

Year 1: To pilot the prototype in at least 100 groups of SHGs of AmritaSREE to get them on board and functional. 

Year 2 & 3: To scale the solution across all the 250,00 SHGs across kerala. Ensure that it meets the requirements of data quality and management, real-time process tracking and monitoring, and customized automatic generation of reports and alerts for the system users. With the data getting gathered at the cloud storage over the time, the data scientists would parallely start working along with the software developers to model and integrate data analytics modules to enhance the application.

Year 4: Further growth to scale across the states to cover other SHGs of the AmritaSREE program. Also we would be collaborating the NABARD supported SHG programs to bring the programs onboard to a common platform. This would set stage for analysis of the different program behaviors there by allowing cross program knowledge sharing and learning.

Year 5: Scaling up by providing value added information to the users and enhancing the features of the data analytics; making it available on the google store to reach word; gathering expert feedback and customer feedback to enable application to other contexts in other countries, providing technical support, demos, and application.

Our institution has over a decade of experience in women’s empowerment programs with this target population. The major challenges we anticipate to face in deploying the project as as below.

1. Gaining the women's trust to port themselves to a digital platform. 
2. Not every woman in a group might have a smartphone to support the mobile application. 
3. For the sub-group that has access to smartphones may not be digitally literate to use the application effectively. 
4. Women in Indian households are usually occupied by multitude of activities and we have often found it hard to gather them together at a particular time slot on a regular basis. 
5. Also the social cohesion varies across the different communities or states making it difficult to have a common deployment plan.

Given below are the plan to overcome the previously mentioned challenges.

1. Women's trust can only be gained by planning proper training sessions along with interesting and interactive training content. Our solution would include required training modules incorporated to the mobile and web applications.
2. In our deployment plan, each of the SHGs would be assigned with one data entry officer to ensure the timely data entry. If required the team will be provided with an android based tablets for the group to use it. The alerts and notifications automatically generated by the system would be sent to even basic mobile phones.
3. Proper training to use the SAS application suite should be provided through the data entry officer.
4. Meeting planner and scheduler is made part of the mobile application so that the women can have commonly agreed time and date scheduled for the SHG meetings. 
5. A group of SHGs forms a cluster and a group of clusters forms a zone. Customized deployment plans would be developed at each hierarchical level, monitored by the staff assigned at the respective levels.

Current size of the team: 6

• Project Lead/ Development Lead - (1) - full time
• Mobile App developer - (1) - Full time
• Web App developer - (1) - Full time
• SHG consultant - (1) - part time
• Data Scientist/ Data engineer - (1) - full time
• Field Deployment Staff - (1) - part time (on demand)

• Team consists of technology experts who are experienced in developing and deploying mobile and web based applications. The software solutions  developed by the team were always motivated by the societal challenges that are left unaddressed.
• The primary delegate of the SAS project is part of AMMACHI Labs that is interdisciplinary in nature. The whole team consists of both Technology group and Social Science group making it easier to design, develop and deploy technology solutions with societal impact.
• Parent organisation having its own SHG program named AmritaSREE that covers 2,50,000 women across Kerala, India. This makes the solution deployment more hassle free.
• Over the past decade, AMMACHI Labs has developed its own skill training model for the rural women population as part of its Women's Empowerment programs. This has been a successful model  incorporating technology based skill training and personality development through soft skills training to every single rural project deployment.

• AmritaSREE (Amrita Self Reliance Education & Employment)
• Mata Amritanandamayi Math

Over the years of hard work there are some SHGs who has managed to stabilize themselves economically by ensuring enough money pool and regular profits. According to our plan, we would be collecting a reasonable amount of revenue from the SHGs that are successful and has the potential to sustain themselves sustainably.

The women Self Help Groups targeted by this project comes from a low-income background. So providing the capacity and support to run sustainable and profitable SHGs inturn directly impacts the betterment of women's life in rural or semi urban population.

SAS application leverages heavily on the data science and data mining domains of information processing and value generation. The objective of the solution is to provide intelligent support to the SHGs, thereby supporting them to undertake best practices that can better sustain the SHG. The outcomes and deliverable of this comprises of 

1. context specific modeling the SHGs behaviors
2. Automatic assessment of the performance of the SHGs.
3. identification and modeling of best practices, 
4. Real time monitoring of the SHG operations to provide timely warning for potential threats.

All the above mentioned milestones aligns with type of problems that can be solved using Artificial Intelligence or Data Mining techniques. 

Potential Use-cases of Artificial Intelligence

Given below are some use-cases of advancing the SAS application mapped to the potential data mining techniques that can be used to achieve them.

• Cluster Analysis to identify SHGs to measure the similarity of SHG behavior. Clustering and exploring these clusters would lead us to behavior modeling.
• Association Analysis of the SHG operations would help us to discover the group operations that are correlated to each other. These correlations would lead us to recommend related success stories and failure stories there by enabling cross group (or cluster or program) learning.
• Exploratory analysis and Statistical tests like distribution tests, hypothesis formation and testing, correlational analysis and multivariate analysis would help us to gain insight about the parameters of SHG performance assessment.
• Time series analysis and segmentation clustering techniques