4 Types of Data Analytics in project-oriented companies

Data analytics?is the process of analyzing raw data and drawing conclusions from that information. In recent years several data analytics techniques and processes have been incorporated into mechanical processes and algorithms that process raw data for human consumption.

Data analytics?is important because it helps businesses optimize their performances. Implementing it into the business model means companies can help reduce costs by identifying more efficient and effective ways of doing business.

Project management?also uses numerous types of analytics and their related techniques in order to describe statuses, find causes and diagnoses, predict, and optimize processes. It can help to improve the quality and also the efficiency of analysis and can lead to time and cost-saving in project execution.

Different Types and Maturity of Data Analytics :

There are different types of data analytics with different maturity levels and applications. For a business case and purpose, certain types of analytics will be selected. The figures below illustrate the main categories of data analytics, ranging from data identification and quality as the initial step up to prescriptive analytics, which is about optimizing processes and systems in terms of efficiency and effectiveness.

Figure 1: Different types of analytics and maturity?

A brief explanation of each of the said types and steps is as follow:?

1. Data Identification/ governance and Quality

Data gathering and classification/grouping are the first essential steps toward any type of data analytics.

Analyzing data involves a few steps as follows:

• Purpose and Objectives: Identifying the purpose of analysis and knowing the mission and goal of that is always the main question that needs to be answered. There should be a holistic view of what is required and to what extent. Data analytics processes could be costly considering their size and purpose so it is very important for decision-makers to consider all aspects of the exercise before and throughout the exercise. ??
• Identifying Data Sources and Characteristics. Once the purpose has been defined, it is time to determine the source of data, its related systems, types, categories, location, and other characteristics.

As part of this preparation, data mining may be utilized to create structured data from unstructured or semi-structured data, and standardization may be applied for the purpose of integration, aggregation, filtering, etc.?????

• Data Quality.

“Be a yardstick of quality. Some people aren't used to an environment where excellence is expected.”

Steve Jobs

Data quality involves activities and assessments to make sure about data’s reliability and fitness to serve its purpose. Some aspects of data quality include completeness, consistency, accuracy, timeliness, relevancy, logicality, integrability, and uniqueness.??

To have quality data, enterprise-level controls and data governance need to be taken in place in any organization including project-oriented companies. That includes defining applicable policies and procedures, using relevant technologies as enablers, standardization, establishing quality assurance and control processes as well as having a robust change control in place.

Moreover, cleaning the data is an important stage that needs to be done in order to have quality data. Data clean-up has a few steps such as removing duplication, irrelevant and unwanted observations from data, identifying and filtering outliers, handling missing data, and validating data. ??

2. Descriptive analytics:?

This describes what has happened over or up to a given period. Business intelligence (BI) tools such as Power BI, Tableau, and other BI tools are used to visualize data in the form of pie charts, bar charts, line graphs, tables, or generated narratives.

Most parts of project reports such as daily, weekly, or monthly reports of the projects usually have elements of descriptive analytics. Projects s-curves, EVM metrics, tabular reports, and project schedule Gantt charts all are samples of such elements.

While descriptive analytics can utilize data from a single source or system, there are times when data models of different types need to be created by integrating data from multiple sources.

Data models consist of some Fact and dimension tables:

Figure 2: A sample data model consists of FACT and Dimension Tables?

In more complex cases the multi-dimensional model may need to be involved. In that case, Datawarehouse and data lakes might require to be developed. They can help to ease access to a wide range of data throughout the organizations.

“A Data warehouse is a subject-oriented, integrated, time-varying, non-volatile collection of data that is used primarily in organizational decision making.[1]”

Datawarehouses are usually subject-oriented. That means data is categorized and stored by business subjects rather than by systems or applications. Below figure shows some other characteristics of data warehouses:

3. Diagnostic analytics:?

This focuses more to examine data to answer the question of why something happened. Diagnostic analytics involves more diverse data inputs and a bit of hypothesizing and involves drill-down, data discovery, and data mining. ?Did the weather affect productivity? What is the impact of different tools on productivity?

Diagnostic analytics in project management can include comparing and examining the performance of different contractors to find the drivers of change, identifying the root cause of defects of welds, and examining Time on Tool to find unproductive times.

4. Predictive analytics:?

This moves to what is likely going to happen in the near term.

Predictive analytics uses a set of techniques that analyse current and historical data to determine what is most likely or (not likely) to happen. It involves:

·????????Choose the right data based on domain knowledge and relationships among variables

·????????Choose the right techniques to yield insight into possible outcomes

·????????Determine the likelihood of possible outcomes given initial boundary conditions

Statistics and mathematical analysis and machine learning techniques usually are the tools that are being used for predictive analytics. ?

Figure 3: some machine learning techniques used for predictive analytics.

Predictive analytics also can effectively be applied in project management in different areas and subjects for example to predict the cost and duration of the activities or for benchmarking purposes.

In order to implement predictive analytics in the projects several steps need to be taken:

1)???Required data identification and data gathering process, data types, and database creation based on the identified domain

2)???Perform statistical analysis to find potential factors and features that may have an impact on the outcome (Features engineering)

3)???Studying what combination of features is most suited to be used in the statistic analysis or machine learning models

4)???Determining the best model to be used for the prediction of the outcome ?

5)???Test the model with the different datasets to make sure about accuracy and effectiveness and make the adjustment if it is required

?Regression analysis is one of the most useful predictive analytics techniques that can be used in various project management domains. However, other techniques such as clustering and classification can also be used.

5. Prescriptive analytics:?

Prescriptive analytics is a?form of advanced analytics that examines data or content to answer the question “What should be done?” or “What can we do to make a certain situation happen?”, and is characterized by techniques such as graph analysis, simulation, complex event processing, neural networks, recommendation engines, heuristics, and machine learning.

This type of analytics also can be widely used in project management in different domains such as quality control, resource management, maximizing productivity, schedule crashing, and fast-tracking.

Major Components of a Project Management Data Model:

In project-oriented companies, a proper data model or data warehouse may include several components. Such data models can be formed by integrating different systems. Below are some major data/systems that can be a part of the plan:?

Figure 4: Major parts of data model/ data warehouse in project-oriented companies?

1-???WBS, Integrated Master Schedule

In the project environment, Work Breakdown Structure (WBS) and IMS should be the core component of the model or data warehouse that may be created for different types of data analytics illustrated in the figure no. 1

Earned Value Management System (EVMS) or Earned Value data should also be a part of the model as they are a major part of descriptive analytics and as well as diagnostic analytics and can be used for predictive and prescriptive analytics.

2-???Enterprise Resource Planning (ERP) system and Material Requirement Planning (MRP).

ERP and MRP contain some essential parts of project data such as financial, accounting, invoices, supply chain, materials demands, and statuses. These are important data to consider for different types of analytics.??

3-???BIM/ 3D model/ MTOs and other technical data.

Construction companies usually use 3D models which are created during detailed design engineering phases and contain essential technical data such as MTOs/ quantities, types of materials, specifications, dimensions, locations of the materials, etc. If a proper integration is created between 3D models and other systems, it enriches the analysis in all types of analytics including predictive and prescriptive. ??

4-???Benchmarking, estimation data, industry, internal productivity data, etc.

While a model with the above specification can be used for the internal benchmarking database it also can be integrated with any existing industry or internal benchmarking database so the comparison between current data and the benchmarks also would be an important capability that can be used. ??

[1] --W.H. Inmon, Building the Data Warehouse, 1992--