The customer manufactures several products related to Industrial Automation. The need was to forecast accurate product demand so right inventory levels could be maintained.
R, Time Series Forecasting algorithms
The products were categorized based on the historical monthly sales
data. Using feature engineering, the missing values were treated first. Using Dickey-Fuller test, the data was tested for Non Stationary. The data did have trends, and seasonality, which was treated and the data was made stationary.
Later, various time series algorithms like ARIMA, Holt-Winters, RNN LSTM were applied to the data. For the products with regular sales, ARIMA gave an accuracy of over 85%, which was fine tuned to get the forecasting accuracy of over 90% using Holt-Winters model. For the products with seasonal sales or non regular sales, RNN LSTM worked very well with a forecasting accuracy of over 95%.
We suggested the customer to use prediction range over absolute predictions, which was accepted by the customer.
The customer’s customer ( An Organization of Govt. of India), wanted to do a Preventive Maintenance Proof of Concept for Air Conditioners. They asked to prepare a report on a set of data generated using various test applied on the Air Conditioners.
R, Data Visualization, K Means Clustering
The supplied data was provided only for 18 hours with details like Voltage, Current, Vibration, Compressor Temperature, Room Temperature, and date time component.
The data did not have a clearly defined breakdown condition. Neither the data was qualified for time series forecasting. It was clearly a case of un supervised learning. We also for relationships between the variables and found that only Current, Vibration and Compressor temperature are good enough to represent the data set. We then applied K-Means Clustering algorithm to the data to create segments with data points which have more or less similar characteristics. With a K-Value of 15, we could distinctly identify on cluster which more or less, had data points, which were defining probable Air Conditioner breakdown condition.
A report on the findings was created with 3D visualizations of the Clusters along with our findings and submitted to the customer. We requested for more clearly defined breakdown conditions which can be used for more accurate predictions.
The customer’s customer ( A Leading Oil Refinery in Saudi Arabia ), required Operational dashboards to track certain KPIs of there multiple plants across the country, near real time.
Power BI, Power BI Report Server, SQL Server Stored Procedures
The Customer wanted to track KPIs for Process Area, Tank Farms, Energy, Environment, Heat, Marine , YRD Corrosion and Feed Products. A total of 12 main dashboards and 8 drill through reports were developed.
The data source was a data generated by machines and sensors. This data was already in a structured form. The customer insisted to use SQL Server Stored procedure to pull the required data. Parameterize stored procedures were developed and used as a data source for Power BI. The dashboard layout were complex, as the customer wanted to see the data in certain manner. Some of the features asked were not directly supported by Power BI. We had to do lot of research and develop a customer solution to address the business problem.
Another challenge was that the customer had on-premise installation. Power BI deployment on cloud is much easier as compared to on-premise deployment. Moreover, the dashboards were required to be rendered through the customer’s intranet portal which also threw some on-time integration challenges, which we successfully addressed and the dashboards were deployed successfully.