Data is the new oil…
For those business owners and Top Management who still have not realized it, they need to be aware that not only have the Steam and Computer Industrial revolutions passed a long time ago, even the Digital Revolution is gone already.We are now in the Fourth Industrial revolution of Cloud computing, Artificial Intelligence and Data. You might not be aware but each day we consume products from data science and machine learning.
- After watching a video on you tube, you receive a recommendation for a similar video based on the previous video you watched, which most likely suits your taste of videos
- Facebook sends you a list of “people you may know” based on your activities on the platform, amongst which many are actually people you know
- Amazon selling items cheaper when competition is high and more expensive when competition is low
- When performing a search in google you get auto complete suggestions in real time, which match with what you were searching for
All the above are examples of how we live in a world today, where you are either embracing data science or completion is embracing it and stealing a huge share of your customers.
The catalysts behind all this data science obsession are: the increase in computing power, increase in data storage capacity, and increase in data volume generated.
As seen in the figure below,there is an exponential increase in the amount of data we are generationg, with the volume of data global data doubling every 2 years.
What Data Scientists actually do
More than 20 years ago, Peter Drucker made the following statement,“in God we trust… everyone else must bring data”
In a company and in life in general, knowledge is power .And we also know that data brings knowledge. Therefore data scientists are supposed to be a key player in the modern industry, using data to bring knowledge and to strengthen the position or performance of a company in the face of its competitors.
Although data scientists come in all shapes and sizes and from various industries and background, they usually possess advanced knowledge in statistics, computer science, information science, operations research and mathematics.
And typically are skilled in data visualization, data mining, software development, machine learning, databases and information management .With these they help companies make better informed decisions and hence improve efficiency and profitability.
Not all these skills are mastered by most data scientists, but with at least a combination of two of these skills the data scientist is ready to help the company make better decisions.
Data scientists typically use open source programming languages like R and Python when performing analytics tasks and typically go through the following steps in their workflow:
- Problem framing: Where the objectives of the analysis are clearly defined
- Data collection: Where the data to be used is sourced .Could be structured or unstructured, and collected either in real-time or in batches
- Exploratory Data Analysis: Where descriptive statistics and visualization are mostly used to discover hidden patterns in the data
- Data Cleaning: Where data is formatted, duplicates are removed or imputed, etc
- Model building: Where the actual machine learning model is built and the model is trained, tested and validated
- Conclusion and Recommendation: Where the findings and summaries are communicated to the stakeholders along with proposed action plans and focus areas.
Importance of Data Science to businesses
The service of a data scientist is of great value to a company .With a competent data scientist working an information based environment, below are the improvements the company is bound to observe:
Improved decision making with facts
About 80% of data available to a business is unstructured (such as coming from comments on Facebook, YouTube videos, company website, etc) and the data scientist uses the mastery of programming languages like R and Python in order to aggregated such data and later use to help predict and prescribe courses of action.
A data scientist is usually management’s trusted advisor and strategic partner to help improve the decision-making processes, across the entire organization. The data scientist captures, tracks, and measures performance metrics and other KPIs to help chose the best course of action.
Improve product appeal
Data science includes competitor analytics, trend analysis ,and market analysis which can help make recommendations on the best distribution channels per product .With this insight, the company will get feedback on product and brand affinity and also information on how to improve current business processes and hence customer satisfaction.
Recruiting the best fit
With the huge amount of data available on social media, job sites, etc human resources can leverage on data science to build models to score the candidates with the best fit for the vacancies in their companies. Candidates who are just impostors and only look good on paper will be easily eliminated in favor of the ones with the best fit for the company.
Most companies now use predictive models to predict future employee engagement, before hiring and also after hiring to predict the most effective training to be offered to employees to boost efficiency.
By combining both structured and unstructured data sources about your customers, the data scientist can be able to create micro segments with specific problems and develop specific products and services tailored to each of these segments.
Application of Data Science in different Industries
There are both general applications of data science across all industries and specific applications for specific industries.
This article will briefly introduce the applications of data science.But subsequent articles will focus on one of the many applications in one of the many industries in which it is being used.
You can also leave a comment below ,to request which application of data science and for which industry do you want a more detailed article on.
A) General Application of data science across all industries
- Customer Profiling: Using Regression and Classification techniques to identify the behavior and preferences of each customer and sell goods and services based on this knowledge.
- Customer Experience: Use Exploratory Data Analysis and Predictive Analysis on data acquired both before and post purchase of your company’s goods & services, in order to discover what makes customers satisfied and improve on the customer experience.
- Churn reduction: Use of regression and classification models to identify key factors affecting churn and build predictive models to predict and reduce churn.
- Reduce costs: Using predictive analysis to determine key revenue drivers and potential costs & risks let the company find the right balance between efficiency and marketing cost.
- Risk Mitigation: Using predictive analysis to identify the potential risks and identify optimal actions to take to mitigate them.
- Forecasting: Using time series and order regression methods to help improve forecasts figures in sales, marketing, finance, etc.
- Sentimental Analysis: Using web scrapping on online data and natural language processing techniques to determine customers sentiment/feedback about a specific promotion or campaign.
- Customer Lifetime Value: Using regression techniques to identify the expected lifetime of a customer and hence the expected life time value. With the objective of using other analytics methods to try to increase the customer lifetime value.
- Natural Language Processing:Mining text gotten from websites,emails,customer text messages,etc in order to obtain key insight such as what the customer is looking for.This is usually taken a step further to create chat bots to respond to simple questions from customers,for example.
- Image Processing: Using deep learning to identify image patterns, perform image classification and help improve decision making.
B) More Specific Application of Data Science per industry
- Finance & Banks
- Fraud detection: Using the pattern detection technique to identifying abnormal patterns in customer’s individual behaviors and sending automatic alerts to an audit team for further investigations
- Improve credit rating:Use predictive analysis to come up with more reliable credit rating scores per customer.This will help reduce loan default risk,for example.
- Genomic Data Science: Analyzing genes to come up with micro segments with similar genetic expressions.This will help when administering drugs,so that genes in same micro segment receive dosage of a treatment which corresponds to their genetic expression.
- Medical image Analysis:Using deep learning to analyse images ,before and after a surgery,for example
- Personalized Pricing:Using regression techniques to determine real time and unique pricing for customers,based on the time of the year, and using other optimizing variables.
- Customer acquisition: Using predictive analysis to study patterns and identify potential customers and hence attract them to come use the services of your company.
- Improving equipment maintenance: Using pattern detection and predictive analysis to be able to identify equipment faults and hence improve the maintenance.
- Avoid power outage: Using predictive analysis to identify signs of an equipment about to breakdown and sending a signal for early action to be taken before it arrives at that stage.
- Production optimization: Using analytics to avoid overproduction of products and at same time leveraging on IOT (Internet of Things) to predict potential problems and resolve them before they affect production
- Improve Monetization: Identifying patterns and preferences based on data aggregated from several sources,in order to come up with models which help improve the monetization of the games
- Improve gaming models:Use of predictive analysis to better models and make games more interesting.
- Market Basket Analysis:Using association techniques to discover what people like to buy in combination with other items. This means products most frequently bought together should be displayed close to each other
- Location based offers: Using GPS information to target customers close to the vicinity of the retail outlets and propose them enticing deals to attract them into the retail outlet.
- Real time inventory:using real time algorithms to stay informed on the stock levels in real time.This system could be linked with the suppliers’ systems which could see the supplier being more proactive in bringing the supplies before some fast selling goods go out-of-stock
- Route Optimization: The use of optimization in determining the best route the delivery trucks should take to reach their destination.
- Recommending goods: Using collaborative and content-based filtering to recommend goods most frequently bought together.
- Identify potential customers: Using analytics and sentimental analysis to identify potential customers.
- Predicting trends: Analyzing the market and sales data over a period of time,in order to identify trends in the market and predict what customers would most likely want in the near future.
- Personalized risk pricing: With profiling,individuals will be charged personalized premiums based on their actual behavior, which may vary as the individual changes his lifestyle.
- 360 Degree Customer profiles:With the availability of data from telematics,smartphones?social media,etc the insurance company can use analytics to aggregate and exploit this data to improve efficiency.
- Waste and Fraud reduction: Exploratory Data Analysis can be applied to help identify where the losses and fraud are highest and decisions taken to reduce them.
- Improve Cyber security: Using data analytics to identify patterns of cyber attacks and take actions to prevent and bring the culprits to book.
- Improve defense system : Heat maps can be used to identify hot zones, also classification and clustering can be used to identify attack patterns and measures taken to counter them.
Do not forget to leave a comment to request which of these a data science applications per industry you would want a more detailed article to prepared for and for.