Interactive Designer Required courses: Principles for data analysis

How do interactive designers develop the ability to analyze data?

Firstly, we should have the consciousness of data collection and analysis, and grasp the source of data generation;

Secondly, get the data, find the correlation between the data, deep dig the intrinsic meaning;

Thirdly, master the basic data analysis methods and apply them in actual combat;

Finally, the analysis results are applied to the follow-up work, and the analysis results are tested.

So, the cycle, the formation of a professional habit, a process of work.

From the usual work, summed up the following several data analysis should pay attention to the principle, here, welcome to clap:

　　I. Clear purpose of data analysis

To analyze a piece of data, you first need to be clear about your purpose: Why do you want to collect and analyze such a data? Only after your purpose is clear, can you be sure of what data you want to collect and how to collect it next. Of course, your goal can be a number of small points, such as: the user on the homepage of what content? How important is the login box on the page? As long as these points are a real problem to be solved, list them down, one to collect data.

The results of your analysis may change the whole project, but with the support of the data, the project or requirements will have a new start or detail adjustment.

　　Ii. Understanding the sources of data and collecting

Establish an analytical framework and prioritize data collection according to the points listed in the objective of the analysis. At the same time, it is necessary to understand how the data is generated and how it is obtained. The data statistics tools that are applied to the work are: gold token, microscope, CNZZ statistics, through these statistical tools can facilitate the collection of data, at the same time, interactive designers to maintain communication with the front end, to understand the methods of data statistics, timely add statistical dimensions, please the front end of the students to help bury the statistics code.

　　Third, grasp the method of data analysis

As an interactive designer, we should master several basic methods of data analysis: contrastive analysis, group analysis, structural analysis, mean analysis and cross analysis. Based on these analysis methods, we can have a preliminary understanding of the status quo, causes and future, and go into further in-depth analysis. Such as: the current situation analysis for the current site or page browsing Click to do a data statistics and hot analysis, you can draw the user's browsing path and focus. The reason analysis focuses on a problem and digs into the answer. Future analysis can be used to communicate with product managers on the planning of later products.

　IV. Communication Analysis Results

Before communicating analysis results, be careful not to use only the information on hand to judge, if the hands of the evidence is not enough to fully reflect the actual situation, the results of data analysis as a decision is easy to make mistakes, especially when looking at a single data dimension. The interaction designer should think ahead, consider the questions that the product Manager may ask, and give a response. Make communication efficient and meaningful.

V. The results of deceptive analysis

The data will be deceptive. One of the most famous examples is The Simpsons paradox. The two colleges in an American college, law School and Business School, were thought to have sex discrimination when they started school.

Law School: (high female enrolment rate)

Business School: (high enrolment rate for girls)

According to the data of the college, the admission rate of female students is higher than that of boys, but in the overall comments, the admission rate of girls is lower than that of boys.

To avoid this, we should group appropriately and adjust the weights of some groups to measure the potential factors that might affect the relationship based on the business.

　Data is not a panacea

Pre-data can be used to tap user needs, the medium-term data can be used to filter the product function, later data can be used to reflect the success of the product. Throughout the process, the data can also be cited as a communication between the product manager and the interaction designer.

However, we have to recognize the fact that the data is not omnipotent. It can not reflect all the problems: in the early analysis may not be able to find innovative breakthroughs or potential demand points, in the latter part of the effectiveness of the validation, often appear to be very convincing. We have to focus on the data with an objective attitude, from different perspectives, and product managers to maintain effective communication.

In addition to the above principles, in the data analysis process, we should also avoid the following situations:

1, the project is urgent, time is not enough

In the early stage of data analysis, first of all to do a schedule, including the following aspects: collection of data, collation of data, analysis of data, summary report. Estimate the time that each content will take, and mark out the key points and arrange the time appropriately.

2, pay attention to collection, analysis is not enough

The focus of data analysis should be on analysis, not on the collection of data. After ensuring sufficient data information should be immediately put into the finishing and analysis phase. If spent a lot of time to collect, in the deadline before the basic no time for analysis, the final submission will only be a superficial summary, and after in-depth analysis of the data report is really valuable.

3, attention to the timeliness of data

Data can tell us what happened in the past (such as user preferences, advertising effects, etc.), but over time, the data will change accordingly. Data is time-sensitive, and too long ago the data may not be able to reflect the current situation, and can not be used to make design decisions. The more real-time the data, the more you can use this data to make the most timely adjustments to the current problem.