SMEs are struggling to integrate Data Science/AI/Predictive Analytics into their daily decision making. This is due to various reasons such as:
- Cannot afford to recruit data science teams.
- It’s difficult to attract the right talents.
- Lack of right-skilled people to lead this kind of projects.
During the COVID-19 crisis this situation will get worse and SMEs will fail to stay competitive to large corporations.
What it does
It will provide a set of predictive analytics tools designed for SMEs. These tools will cover the basic needs for these enterprizes on forecasting, segmentation & churn risk. The initial tool will focus on forecasting. The main idea is to utilize various forecasting algorithms (traditional forecasting methodologies & machine learning algorithms) together with anomaly detection algorithms and serve the business users with optimized forecasting accuracy, regarding their revenue, cash flow, product demand, web pages performance etc.
How I built it
The prototype app is built by using the R Shiny framework along with shinydashboard & various other R libraries. Shiny allows you to build interactive web standalone apps straight from R, using the Bootstrap framework. The final prototype is hosted on https://www.shinyapps.io/ which is a dedicated platform for hosting this kind of applications. The compute resources will scale as the number users increases.
Challenges I ran into
It was difficult to follow a time plan in order to finish all tasks on time. Especially the video creation was a really difficult, demanding and long process...
Accomplishments that I'm proud of
SMB Analytika will help small and medium-sized enterprises to integrate predictive analytics/data science/AI in their daily decision making. All tools are designed to help an average business user to start using these in a few minutes, without any official training.
What's next for SMB Analytika
The tools will be designed in such a way that average business users, with minimal training & expertise, will be able to use these in a matter of minutes. Future tools will include customer segmentation, churn risk modelling. All these tools can be turned into ready to use products in the next few months with the appropriate resources.
Although it is early to set a price level, from my experience I could say that a monthly price of 100-500 € per enterprise is feasible. Of course, it will vary based on the number of tools, their complexity, the support etc.