Data Analyst (DA) Salaries and Rates in 2024

Data Analyst (DA)

A Brief History of Analytics

A simple definition would be ‘the study of analysis’, while a contemporary, and probably more useful definition would say ‘data analytics’ is a tool for business insight and personalised answers to customers.
‘Data analytics’ – now often shortened to ‘analytics’ – has been a crucial component to all types of organisations in recent years. The process of data analytics has evolved and broadened over the years to become very useful.

Lastly, analytics in business can be traced as far back as time management exercises initiated by Frederick Winslow Taylor as early as the 19th century, and in Henry Ford’s measurements of how fast assembly lines should go. Analytics received much more attention in the late 1960s when computers began to emerge as decision-making support systems.

After big data, data warehousing and cloud technology became viable prospects, we saw new forms of data analytics emerge. Data analytics is the research, discovery and interpretation of patterns in data. The newer forms of data analytics include:

• Predictive Analytics
• Big Data Analytics
• Cognitive Analytics
• Prescriptive Analytics
• Descriptive Analytics
• Enterprise Decision Management
• Retail Analytics
• Augmented Analytics
• Web Analytics
• Call Analytics
• Statistics and Computers

Data analysis is based on statistics. And it is said that during the reign of ancient Egyptians, the construction of the pyramid was based on statistics. The state governments in the world have carried out statistical survey based on household registration census and use it in their various plans for work such as taxation. After collecting the data, the purpose of finding information and summarising the relevant information will be carried out. For example, judging the density of a new hospital according to the growth of population in county and city.

Thanks to the development of computers and the evolution of computing technology, the process of analysing data has undergone a radical improvement. In 1880, the information which was collected by the U.S Census Bureau had to wait much longer than seven years before finally being processed and the results published. To help speed the tabulation process, the inventor Herman Hollerith came up with the “tabulating machine”, used in the 1890 census. This device could processes sensible information recorded on punch cards. With this device, the 1890 census was finished in 18 months.

Relational Databases and Non-Relational Databases

It was in the 1970s when Edgar F. Codd first invented relational databases, but it wasn’t until the 1980s they became popular. Soon after relational databases, users wanted to be able to write in sequel (SQL), or ‘structured query language’, to query their database and retrieve data on demand.
RDBM and SQL give us the capability of analysing data on demand, and these technologies are still used quite frequently today. They are easy to interface with, and great for keeping accurate records.
However, on the other hand, relational databases are designed to be very unforgiving and not meant to translate unstructured data.

The internet went truly mainstream in the mid-1990s, but relational databases couldn’t cope with the crazily expanding data flow, accompanied by large and incompatible data types coming from a multitude of sources. The result was non-relational databases – so called NoSQL. A NoSQL database translates data into multiple data languages and formats very quickly. NoSQL also frees up from SQL’s rigid data organisation by replacing it with a looser, more flexible, ‘disorganised’ storage.

The development of NoSQL was then followed by another series of changes in the internet. The founders of Google, Larry Page and Sergey Brin, designed their search engine to search a single website, and to process and analyse big data by distributed computers. Google’s search engine can return relevant results at visitor’s demand in just a few seconds. The major interests of the system are thus scalability, automation and high performance. A 2004 white paper on MapReduce drew many edges of interests from many engineers and attracted a big number of talents to focus on the problems of processing big data (data analytics).

Data Warehouses

By the late 1980s, the rapidly growing volumes of data were aided by the dramatic decline of hard disk drives. During the digital revolution, data warehouses were conceived to help the transformation of data fed by operational systems into decision support systems. Data warehouses are usually part of the cloud or part of the organisation’s mainframe server. Unlike a relational data base, a data warehouse is usually normally optimised for a rapid response to queries made. Typically, in a data warehouse, tables are left intact with data often timestamped and operation commands such as DELETE or UPDATE less frequently used. If all of the sales transactions were timestamped, an organisation could use a data warehouse to compare the sales trend of each month.

Business Intelligence

The term business intelligence (BI) was first used in 1865; but the current usage, defined by Howard Dresner at Gartner in 1989, described making better business decisions by searching, gathering and especially analysing the data that an organisation had been saving by its computers and other devices. Using the term ‘business intelligence’ to describe the use of technologies to help make business decisions was another step, but it was a bold step that showed great foresight. Large businesses had first used BI to analyse data about customers professionally, as they began to produce data at a rate and scale that made decisions more efficient.

Data Mining

Data mining started in the 1990s and involves the discovery of patterns from large data sets. Results of data analyses that had previously seemed counterintuitive eventually proved to be positive. Quite logically, the occurrence of data mining is closely connected with the development of database and data warehouse technologies. New technologies enable organisations to collect more data and analyse it at the same time, at lower costs and in shorter timeframes. New ways of data analyses, too, have been developed. Firms began to anticipate customer needs on the basis of analysing their historical purchase records.

But data is often misinterpreted: the consumer in the trades who bought two pairs of blue jeans online in the past two years is unlikely to want another pair for at least two or three years. To advertise blue jeans to this person is a waste of time, and an annoyance to the customer.

Big Data

The term big data was first assigned in 2005 by Roger Magoulas, to describe a large volume of data that was thought to be close to unmanageable using the small Business Intelligence tools available at that time. Hadoop was released in 2005, which enabled the processing of this deluge of data — it was built upon Nutch, the world’s first open-source search engine, and incorporated MapReduce, a programming model invented at Google.

The Apache Hadoop software framework, for example, is an open-source software enabling the distribution of processing among nodes (or computers) across a network and the parallel processing of data at scale. Data could be structured or unstructured, streaming in via almost any digital device. Apache Hadoop, and its siblings in the family of Apache open-source frameworks, is made to process big data. A whole new family of open source projects emerged to address the challenge in the latter part of the 2000s, including Apache Spark and Apache Cassandra.

Analytics in the Cloud

In its first incarnation, the term cloud was used to describe the ‘empty space’ between user and supplier. Then in 1997, Ramnath Chellappa, professor at Emory University in Atlanta, suggested that cloud computing was ‘a new computing paradigm where the boundaries of computing will be governed by economic rationale rather than technical limitation’.

One of the first examples of doing cloud computing right was Salesforce in 1999. It is primitive compared with today’s cloud computing, but Salesforce was attempting to exploit the idea of delivering software programmes over the internet. Programmes, or applications, were accessed or downloaded by anyone with access to the internet. A company manager could buy software at a low cost, on-demand, without having to leave the office. As more businesses and organisations understood what the cloud does and how it can be used, it grew in popularity.

The cloud we know today had its ‘baby steps’ back in 1999, when users ‘rent’ its services instead of owning it. Thus, IT vendors eliminate the tedious maintenance work such as trouble-shooting, backups, administration, capacity planning and maintenance. In addition, the cloud is easy and efficient for several business projects, as it is now equipped with enough storage space, has the availability to multiple users, and is capable of handling multiple projects.

Predictive Analytics

The process of analysing data in order to determine trends and patterns is known as predictive analytics. Predictive analytics leverages a number of statistical, modelling, data mining, artificial intelligence, machine learning techniques to make predictions about the future based on data. The predictive models can be used to analyse both current and historical data to understand customers, purchasing patterns, procedural problems and to predict any potential threats and opportunities for an organisation.

Predictive analytics first started in the 1940s, as governments started to use the first computers. Though existed decades ago, in recent years predictive analytics has turned into a concept whose time has come.
As more and more data are put into storage, companies are using predictive analytics to boost their profits and improve their business competitiveness. The vastly growing amount of data stored and the increasing willingness to use data for Business Intelligence, have promoted the use of predictive analytics.

Cognitive Analytics

Most organisations handle data in unstructured form. It is impossible for humans to make sense of unstructured data. Cognitive analytics combines multiple applications to provide context and answers. An organisation has the opportunity to gather data from multiple sources. With cognitive analytics, the unstructured data can be examined, giving the decision-makers a bigger perspective about the internal process, customer preferences, and customer loyalty.

Augmented Analytics

According to Gartner, augmented analytics is Business Intelligence (and insights) ‘automated by combining NLP [natural language processing] and machine learning’ and augmented analytics automates data preparation and data sharing. Both augmented analytics results and access to tools and data are clear, and managers and researchers can engage in decision-making on a daily basis with a high degree of certainty. The researcher and the manager can both see the analytical and numerical results of applying statistical tests, and they both can draw conclusions from this analysis. Once you’ve augmented your analytics, you’ll be ready to make a decision and take action.

At the end of the day, augmented analytics automates all the steps performed by data scientists for having insights and business intelligence. A augmented analytics engine will automatically analyse the organisation’s data, perform data cleansing, conduct analysis and produce insights for business executives or salespeople.

Portfolio Analytics

It represents a portfolio analysis, which is typically what a lending agency or a bank would have, namely, a table of accounts where the return on the loan and the risk for defaulting are different (and often vary) in the present and in the future. These factors could include information that the lender has about the client’s social status (poor, middle class, rich), their location, and other criteria. Portfolio analytics can help the lender balance the return on a loan versus the risk for default. The risk of any particular loan will be determined by factors such as incoming, success of previous loans, or declaring bankruptcy.

HR Analytics

Now referred to as HR analytics, ‘people analytics’ involves behavioural data that can be used to quantify how people work and how they impact organisational management. HR analytics has also been referred to as workforce analytics, talent analytics, talent insights, people insights, colleague insights and human capital analytics. HR analytics has been used to guide business management of human capital, and also used as a strategic tool for creating labour-market trend analytics and forecasts.

Customer Journey Analytics

The customer journey deals with the complete journey of customer that all the customer goes through while interacting with an organisation or brand. Therefore, it records the entire experience of the customer, not just a part.

Customer journey analytics (analysing customer data after it’s been recorded) helps people understand the customer experience – often in real time – and thus can influence the design of the customer experience. It allows for a systematic method of evaluating and monitoring the customer journey and improving that journey. Ultimately, designing and delivering the best customer experience is what we’re after.