Nowadays, almost every aspect of our lives generates data. Gathering and analyzing data to help better understand our customers and businesses is an unbeatable competitive edge. Those insights would help us design better products and services, forecast the market trends and support our decision making to seize new opportunities. Explore and leverage data analytics with us to drive higher return and be more competitive on this volatile market.
Kepro offers data software platform consultancy, implementation & development, enablement / training, and data scientist service.
GENERAL BUSINESS USER
Data are always massive, it is hard to read and understand the meaning in seconds. Data visualization is the graphical representation of information and data. By using visual elements like charts, graphs and maps, it provides an accessible way to identify and understand trends, outliers and patterns easily and quickly. Thus, making data more understandable.
Data visualization is a way of communication when describing our findings from data. A good visualization can present several ideas in a single dashboard. With user interactions, we can obtain their information with interactive dashboard. It also prevent user to look up different reports to obtain the same idea.
Some reports are repetitive, such as financial reports, inventory reports, and etc. With visualization tool, we can automatically update the latest data without any further effort. In addition, dashboard can set to be role based which means user can only access with their information within the same dashboard. This would save a lot of time to prepare different reports for individual roles.
Nowadays, business users have a lot of reports to read. Some reports’ functions just require to monitor the status or make sure the value does not drop out of the confidence level. Visualization tool can help to monitor this kind of process automatically by setting different kind of rules. Once the event trigger, user can receive email acknowledgment and receive the link to access on particular dashboard directly. It saves users a lot of time on finding out the related information.
Most of the raw data are not cleansed which mean that it is not directly processable from machine. Moreover, useful information are always come from difference applications or sources, data integration would be needed before analytic process. Some of the data fields have to do different kind of calculations in order to meet with the business definitions and logics. All of the above kinds of process are to make sure the data are cleans, integrated and ready for use, we conclude as Data Preparation.
In order to enrich with the outcome, we would also apply modeling on the existing data. Like,
- Time-series model to predict the forecast of the future
- Regression model to segment similar individuals into group
- Decision tree and random forest model to predict the outcome
Data Preparation can be done by programming language, however Data Preparation tools can provide business users a better solution to handle with their data need. For example,
- A graphic user interface to present the data operation & flow clear
- Data flows are created by drag & drop instead of writing code
- Easy to understand the logic, apply change and debug
- Data flows are highly reusable, share and collaborate
- Save development time and process automation
Business users are usually the data owner and have the most understanding within the company. Data Preparation tool provide a self-serve environment and no need to reply on other party dependency for creating data analytic flow. It would save a lot of turnaround time and cope with timely market opportunities.
Certainly, we should have defined our destination before the data journey. Business users find out more data are available but how could they utilize with them? In alternative, we have a clear analytic goal, however do the exiting data fulfill the need? Can we obtain external data to fill the gap? In another scenario, I have a lot of data and want to explore hidden insights. Do we have enough time and skills to curate them? All these are challenging questions to business users, however, if we make good use with the data, huge return will be obtained.
With more business users can easily access and make use with the data, data security & control, management are important issues for IT to encounter. For some industries, there will be different laws and compliances for company to cope with in order to prevent leakage or disclose unnecessary personal data to other parties, like GDPR.
Data Governance platform would provide data processing and metadata management. Not only it could let business users to access with the data, it would also let them know easily where their required data are. We can also assure the quality of data and be the single point of truth for whole organization. In addition, Stewardship flow to ensure the corrective records are amended or changed. For sensitive data like personal information, platform would perform data masking to disclose. For data governance approach, data accessibility, availability, integrity, usability and security are assured within the system by set of rules and policies regarding the data, granting or restricting access to data as needed.
In order to fulfill the service level of data availability, a well design data architecture have to be considered beforehand. Technologies like Data Lake, Data warehouse, RDBMS/NoSQL, ETL process, streaming engine, APIs, Cloud Services, Securities issues are all putting in combination of considerations. Moreover, it will involve a lot of process, like, find out the project sponsors, discuss the benefits & ROI with the stakeholders, requirements from critical users, applications owners, etc. We may require assistant to analyze the pros and cons on different methodology and find out the most optimized solution for the whole company.
Company must have data sharing between applications, in particular, legacy application with SAAS. In a long run, we may develop many point to point API connections which will become less control and a mess. An API platform would help to address this by control the use of the APIs. We can know the loading of each API and provide more resource if necessary. We can also build a better API design and architecture instead of point-to-point connections. Those API will be highly reusable, utilize CI/CD model to fast releasing new features and bug fixes to keep users or customers better experience.