Without denial, Data Science has brought a paradigm shift in the way business use to happen. Be it B2B or B2C, business have come much closer to their customer than before. It won’t be wrong when I say that each one of us experience some or other kind of data science algorithm daily in our lives. Be it your facebook recommendation of people you would like to add, or the SMS campaign you receive daily or customised offers from our banks, telecom service provider or online shopping portal. Almost daily, some or other machine learning algorithm decides our day. And why only life revolving around online medium, even our offline lives is also not left untouched by such algorithm. From the kind of offers we encounter when we walk into a retail store or a billboard advertisement we see on the roads or print advertisement on magazines/ newspaper, all these may be result of some or other kind of statistical model. However, all these attempts of reaching the customer efficiently and effectively is confined to corporates and retail giants who are crunching numbers in petabytes to understand our buying behaviour. Amidst all these, there is one section of society that has been largely ignored: the poor and the deprived.
On the hindsight, it seems there is not much to do when it comes to the bottom of the pyramid population but then if one digs deeply and combines consumerism with policy making, the entire Pandora box would open.
Analytics techniques could be leveraged to come up with strategies for the welfare of the poor and needy.
Conjoint analysis, the most widely used quantitative method, is used to measure preferences for product features, to learn how changes to price affect demand for products or services, and to forecast the acceptance of a product if it’s taken to the market. It includes surveying the poor and the needy what they prefer in a product, or what attributes they find most important to buy them. Moreover, conjoint analysis employs the more realistic context of respondents evaluating potential product profiles ranging from simple mobile phones to tractors. The pricing of the products/services could be moderated by understanding the preferences of the destitute. The results of such analysis can be considered and incorporated into the state policies such that the prices of the products sold to the impoverished will not be augmented with addition of fruitless fanciful features. To make the essential amenities affordable to the poor, analytics techniques could be of great help.
Natural Language processing(NLP) is looked at as combination of computer science, artificial intelligence, and computational linguistics concerned with the interactions between computers and human (natural) languages. NLP is human –computer interaction. Sentiment analysis employs NLP for mining and extracting the attitudes and opinions of people across the blogs, articles, reviews and social media portal. It is widely used to discover how people feel about a particular subject. This would prove effective when applied to understand the opinion and the sentiments of the downtrodden towards a say a loan waiver plan or a development and a relief scheme. Their needs could be catered accordingly if this sentiment analysis technique is utilized.
Network analysis is a phenomenal technique to understand how social relationships influence the purchase behaviour and loyalty of people. Though this can be effectual in improving campaign effectiveness and retention of high-value, this technique can also be used to understand the way people are connected in a village or small town and how they influence each other’s decisions. Who are the influencer and who has highest closeness or between ness or hold high Eigen vector value. Thus the information about the influencers can be used to promote healthy lifestyle and behaviour among the impoverished. This works the best in educating the poor by bringing an awareness among the people on deadly communicable diseases, social responsibilities, the merits of social activities. This can be further extended to educate them on Women empowerment, prevention of anti-social crimes like theft, rape and trafficking. Self Help Groups and Rural Employment schemes can be created through the influencers completely towards the upliftment of the society altogether.
Predictive modelling is another powerful technique that can be used for fatal disease management in rural areas. Moreover, predictive modelling techniques provide with the strategies for stopping the spread of deadly diseases. Effective strategies can be obtained for stopping the spread of deadly diseases through building a realistic contact network, predicting the spread of disease through the network, and quantifying the impact of intervention. Also, understanding of the frequency of each degree and the ways and means of disease transmission in the population can be successfully achieved with Predictive modelling. The Probability that an infected individual will transmit the disease to another individual can aid in predicting epidemics based on socio-economic parameters, lifestyle, data from local health clinics and pharmacy outlets, geographic and weather conditions and demographics. A contingency as well as a mitigation plan can be created well in advance by analysing and interpreting the results of prediction. Network analysis coupled with Predictive modelling can thus be used to educate and warn the people about upcoming threats and also promote remedies and precautions toward the fatally communicable diseases. This would promote a healthy living among the impoverished there by raising their standard of living.
Classification technique proves useful to classify people into variegated social classes. Sub classification of poor into smaller classes can be done in order to customise poverty eradication plan for each class. The needs and the preferences of each class of the society would vary from one another. Not everyone requires loan waiver, not ever one requires subsidy, not everyone needs farms loans and not everyone needs health insurance schemes. Classifying people into the right sets of groups and then making customised poverty eradication plans (bundle of few plans considering the behaviour or attributes of the group) would not only make the plan effective but also would make the right plan reach the right set of audiences.
Skill development plans and methodologies can be based on various statistical models that talk about how to teach, when to teach, whom to teach and finally what to teach rather than teaching everything to every one at a blanket level. Not everyone would be inclined to learn or know sewing or farming or any other activity for that matter. The inferences from the statistical modelling techniques that were discussed earlier can be collated to derive fruitful insights to identify and map the needs and the interests of the destitute and the downtrodden in any set up. This would definitely instil in them a sense of responsibility to work hard and earn and thus to raise their standards of living. Similarly, the primary education in villages show how handicapped our students are. It is really distressing to see when an eighth grader can’t read or write or perform a simple mathematical calculation. The quality of education can be improved by classification technique based on machine learning algorithms that would help identify the pain points and thus recommend a solution. Even Artificial Intelligence techniques play an important role in determining the weightages of the factors that have contributed or affected the current education level of the students from the downtrodden society in rural areas.
Data is required to do all the aforementioned analysis to draw meaningful insights on an issue and make smart decisions. Big data, which is the buzzword today, offers for example the possibility to predict food shortages by combining variables such as drought, weather conditions, migrations, market prices, seasonal variation and previous productions. Kenya is one of the pioneers in Africa regarding opening up their data. As the World Economic Forum report says that in 2009 they opened the Open Data Portal where the government shares 12 years of detailed information regarding their expenditures, household income surveys as well as health facilities and school locations. The portal can be accessed by anyone via the web or via mobiles. This could serve as a good data source for doing analytics to understand the current status of different social groups and thus frame strategies to eradicate poverty.
–Sray Agarwal, Chief Manager-Business Analytics,
TimesPro (Times Centre for Learning Ltd.)
Try deep learning using MATLAB