Enterprise Intelligence and Analytics Recommend Strategies to Mitigate

Question :

First, research and explain each of the Data Quality characteristics For each one, what is it and how is it measured?

Then, identify the issues with the Gaudian Pacific data provided based on the quality characteristics. State whether the data fits each characteristic, and why (justification).

Finally, research and recommend strategies to mitigate issues with each characteristic (whether this data has them or not).

Answer :

1.1.1 Rolls Royce

In 1904, the owner Henry Royce made the first car. In the same year he met Charles Rolls who was a well-known car seller in London. The duo agreed that Royce would build cars which would be sold by Rolls exclusively. Thus, Rolls Royce was formed. Silver Ghost was one of the early milestones of the company (Rolls Royce, 2020). From his first aero engine, Eagle to his first transatlantic flight, Royce created multiple records within half a century.  In 1971 car business was separated from the aero engine business. The motor division was acquired by Vickers PLC in 1971. After many ups and downs, milestones and business structure changes, Rolls Royce is now a leading power provider in land, water and air vehicles. 

1.1.2 Tesco PLC

Tesco PLC is the biggest name is the British Retail Industry. It began as a grocery stall business in 1919 by Jack Cohen, in East London. The first Tesco store came up in north London in 1929. Due to the initiatives taken by Cohen, the British Parliament passed the Resale Prices Act in 1964, which eliminated Retail Price Maintenance and it helped the small retailers to get the benefit of economies of scale. By 1970s the company was already an owner of hundreds of grocery stores in the country. From its first petrol station store in 1974 to being the largest petrol retailer in the country in 1991, it was contributed towards 12.5% of the total petrol sold in the UK. Later in 1995, it became the first supermarket to offer loyalty programs to customers. It began its online presence in 2000 and community services in 2015 (Tesco PLC, 2020). In 2018, Booker Group was merged with Tesco which further increased the brand’s market share and global presence. 

1.2 Big Data Implementation

1.2.1 Rolls Royce

Rolls Royce operations can be divided into three major areas which are, designing, manufacturing and after-sales services. During design simulations huge terabytes of data are processed through highly advanced computers which helps the company decide on the fate of the particular design. The company focuses on the quality of the data more than the quantity. The company has adopted Internet of Things solutions to process the generated data. One fan blade of the company in manufacturing simulation produces half a terabyte data which amounts to three petabytes for a single component annually (Marr, How Big Data Drives Success At Rolls-Royce, 2015). The generated data from operations amalgamates information which helps in detecting engines in need for maintenance. The Ship Intelligence Initiative is an example of the implication of Big Data based services. Under this technique, personnel are automatically scanned, automated piloting is enabled, and sensors detect issues. Its Engine Health Management System is one of the examples. Through this system data is collected from all the components of the engine and allows real-time monitoring in any part of the air or water. There is a central control room in Derby which manages this data. The company uses the Microsoft cloud services for its Big Data Analysis. There is no specific mention of a particular tool but it’s highly probable that the company uses Azure. It is a Big Data and Machine Learning platform that simplifies data analysis. 

1.2.2 Tesco PLC

Tesco Labs division is responsible for the Big Data Analytics at Tesco. The company’s real time data analysis with the help of Teradata and Hadoop enables it to gain insights on customer behaviour, increase supply chain efficiency, and reduce wastage. The key strategy of the company is to combine the functions of Teradata and Hadoop (Marr, Big Data At Tesco: Real Time Analytics At The UK Grocery Retail Giant, 2016). The company uses cluster based analysis with more than 100 million data points. These data points connects more than 3000 stores in UK and overseas stores of Thailand and India as well. These data points are the results one time tracking only. The focus is on data storage and not batch-wise movement of data. Next, the company is using Hadoop to create a Tesco Data Lake Model which has a centralized control and cloud based storage. The data is coded for real-time usage. It also uses sensor data to control the freezer temperatures across stores. The technique used is the predictive algorithm which automatically sends update about maintenance requirement. Tesco also uses Github which allows systematic coded storage. Through the use of these technologies, Tesco is able to select the kind of data that needs to be provided to the different stakeholders.    

1.3 Summary

From the above discussion, it can be understood that both Rolls Royce and Tesco efficiently uses Big Data Analytics and both companies have the vision to improve the system for improved services. The common feature in both the cases was the use of sensors to detect issues. While Rolls Royce have been using Big Data in different forms before it became a trend, the system is comparatively new for Tesco. Therefore, while Rolls Royce is already mastering the art, Tesco is still in the learning and development stage. Two interesting points that emerged from the case studies are the use of real-time analysis by Rolls Royce mid-air or mid-ocean and the centralized freezer control system at Tesco. 

2 Part B: Guardian Pacific Sales Data

2.1 Data Quality Characteristics

Following are the data quality characteristics:

• Accuracy: It refers to the correctness of the data. Data that is error free and the measured figures match the true figures represent accurate data. Error ratio and deviation metrics can be used to measure it. 

• Completeness: It refers to the quantity of data that is enough to form inferences. In a complete set of data there is no incomplete or unfulfilled figures. The percentage of the available records can be used to measure completeness (Scannapieco, Missier, & Batini, 2005). 

• Reliability: The source from where the data is collected represents its reliability. There should be no contradiction between multiple sources. Cross-checking of the sources would ensure reliability. 

• Relevance: The data should be meaningful for the target audience. This means the audience should be able to use the data for their benefit. There is no true measure of relevancy but prior research makes the data relevant. 

• Timeliness/Currency: It refers to the timely collection of data. This ensures the ease of data tracking over a period of time. Time variance metrics is the most useful measurement tool. 

• Uniqueness: It refers to the reduction of redundancy in data. Unique data means data that has not been repeated or duplicated by mistake. Percentage of duplication is used to measure this characteristics. 

• Validity: It refers to data integrity. The similarity of data points and the use of the appropriate format represents valid data. Again, the percentage of data which complies with the required format is a measure of data validity (Sidi, et al., 2013). 

• Consistency: It refers to uniform data sets. There should be zero conflict in the prepared data so that data from different sets are identical. The mathematical tools of range, standard deviation and variance is used to measure consistency.

2.2 Issues with Guardian Pacific Data

Four datasets have been provided by Guardian Pacific, International Sales, Company Funding, Insurance and Residential Sales. The international sales data has been provided for 2009 along with the account creation dates. However, the main inconsistency issue in all the sets is with the timeline. Company funding records as early as 2000 is not relevant anymore. The insurance records cannot be verified for any of the data qualities because no dates are available. The residential data sales are from 2008 and so it cannot be matched against international sales for 2009. As cross-checking of sources is not possible, data reliability cannot be measured. 

2.3 Recommendations

There should be data quality dashboard which would reflect the percentage of quality compliance of all the characteristics at any point of time. There should be no manual system for the data entry. This would mitigate the problem of accuracy by eliminating human errors. Data filtering system can be used to remove duplicate data (Scannapieco, Missier, & Batini, 2005). This should be a system which either eliminates the redundant data or merges it with previous similar entries. Metadata should be managed to ensure timeliness of the data. This will prevent obsolete figures. Data quality thresholds ascertain consistency and accuracy. There should also be a specific format of all data categories.   

3 Works Cited

Sidi, F., Panahy, H., Affendey, L., Jabar, M., Ibrahim, H., & Mustapha, A. (2013). Data quality: A survey of data quality dimensions. 

Tesco PLC. (2020). About Us. Retrieved from tescoplc.com: https://www.tescoplc.com/about/