Commonly used data analysis methods

Commonly used data analysis methods are: 1. Comparative analysis method; 2. Group analysis method; 3. Structural analysis method; 4. Retention analysis method; 5. Cross analysis method; 6. Funnel analysis method; 7 , Matrix analysis method; 8. Quadrant analysis method; 9. Trend analysis method; 10. Index analysis method; 11. Comprehensive evaluation analysis method. The "comparative analysis method" is to compare data to analyze the differences between data, including static comparison and dynamic comparison.

1. Comparative analysis method

is the comparative analysis method, which compares data to analyze the differences between data, including static comparison and dynamic Compare. Static comparison is also called horizontal comparison, which is a comparison of different indicators at the same time; dynamic comparison, also called vertical comparison, is a comparison of indicator values ​​in different periods under the same overall conditions. The purpose is to reveal the development, changes and regularity of the things represented by the data.

Horizontal comparison: Horizontal comparison is to compare with yourself. The most common data indicators need to be compared with the target value to answer whether we have achieved the goal; compared with our last month, to answer how much we have grown around the north.

Vertical comparison: To put it simply, it means comparing with others. We need to compare with our competitors to answer our question about our share and position in the market.

2. Group analysis method

Combined with the contrast method, separate objects of different nature in the population and compare them to understand the inherent data relationship.

3. Structural analysis method

Also known as proportion analysis, it analyzes the proportion of each component in the whole and the changes in composition, so as to grasp the characteristics and changing trends of things.

4. Retention analysis method

The retention analysis method is an analysis model used to analyze user participation and activity level. It examines how many users who performed the initial behavior will perform subsequent actions. Behavior. From a user's perspective, the higher the retention rate, the better the product grasps the core needs of users, and more active users will be converted into products, which ultimately helps the company.

For example, we can observe the user retention situation in different time periods and compare the subsequent retention changes of users across various channels, activities, and key behaviors to discover the influencing factors that improve user retention rate, such as observing coupons received. Is the user retention rate higher than that of users who have not received coupons?

5. Cross analysis method

is the three-dimensional analysis method, which is often used to analyze the correlation between variables. A method of cross-displaying data from different dimensions and conducting combined analysis from multiple angles.

Comparative analysis includes both horizontal and vertical comparisons. If you want both horizontal and vertical comparisons, there is the cross analysis method. The cross analysis method is to cross-present data from multiple dimensions and perform combined analysis from multiple angles. When analyzing app data, it is usually divided into iOS and Android. The main function of cross analysis is to segment data from multiple dimensions and discover the most relevant dimensions to explore the reasons for data changes.

6. Funnel analysis method

Combined with comparison and group analysis methods, you can compare the before and after optimization effects of the same link, the conversion rates of different user groups, and the conversion rates of similar products in the same industry. It reflects user behavior status and user conversion status at each stage from the starting point to the end point. Two commonly used indicators are conversion rate and churn rate.

7. Matrix analysis method

is the matrix correlation analysis method, which uses two important attributes of things as the basis for analysis to perform classification correlation analysis to provide a reference for problem solving and resource allocation.

8. Quadrant analysis method

Divide each comparison subject into four quadrants based on different data. If IQ and EQ are divided, they can be divided into two dimensions and four quadrants, and each person has his or her own quadrant. Generally speaking, IQ guarantees a person's lower limit, and EQ increases a person's upper limit.

An example of the quadrant analysis method used in actual work before. Generally, registered users of p2p products are attracted by third-party channels. If the quality and quantity of traffic sources can be divided into four quadrants, then a fixed time point is selected to compare the traffic cost-effectiveness of each channel. The quality can be measured by the total amount of retention. as standard. Continue to maintain high-quality and high-quantity channels, expand the introduction quantity of high-quality and low-quantity channels, pass low-quality and low-quantity, and try the delivery strategies and requirements of low-quality and high-quantity. Such quadrant analysis allows us to conduct comparative analysis. You get a very intuitive and quick result.

9. Trend Analysis Method

When there is a lot of data and we want to discover data information from the data faster and more conveniently, we need to use the power of graphics at this time. The so-called graphics The power is to draw it with the help of EXCEl or other drawing tools.

Trend analysis is generally used for long-term tracking of core indicators, such as click-through rate, GMV, and number of active users. Generally, a simple data trend chart is made, but just making a data trend chart is not analysis. It must be like the above, what changes in the trend of the data, is there any periodicity, is there an inflection point, and analyze the reasons behind it, regardless of Is it an internal reason or an external reason. The best output from trend analysis is ratios. There are month-on-month, year-on-year, and fixed-base ratios. For example, how much GDP increased in April 2017 compared with March, this is the month-on-month ratio. The month-on-month ratio reflects the recent changing trend, but it has seasonal effects. In order to eliminate seasonal effects, a year-on-year calculation is introduced. For example, the GDP growth rate in April 2017 compared with April 2016 is the year-on-year growth rate. The fixed base ratio is easier to understand. It means to fix a certain base point. For example, the data in January 2017 is used as the base point. The fixed base ratio is the comparison between the data in May 2017 and the data in January 2017.

10. Indicator Analysis Method

In actual work, when we get some visual data charts or Excel tables, we can directly use some basic indicators in statistics to do data analysis. , such as average, mode, median, maximum value, minimum value, etc., below we will introduce them respectively:

mean

The average, also called the average analysis method, refers to An analysis method that uses the method of calculating averages to reflect the general level of a certain quantitative characteristic of the population under certain conditions at a certain time and place. Commonly used indicators of the average analysis method include arithmetic mean, harmonic mean, geometric mean, mode and median, etc. The most common one is the arithmetic mean, which is what is commonly known as the average or average.

The average indicator can be used to compare the degree of difference between similar phenomena in different regions, different industries, different units, etc., which is more convincing than using total indicators. In addition, using average indicators to compare changes in certain phenomena in different historical periods can also better explain the trends and patterns found.

Mode, median

The mode is a representative number in the data, which reflects the degree of concentration of the data. For example, the best, the most popular, and the most satisfactory are all related to the mode. Essentially, the mode reflects the data indicators that occur most frequently in the data. When doing data analysis, we can extract some common characteristics of these data indicators, then refine and summarize them, and then draw some suggestions for improvement. .

The median mainly reflects the central tendency of a set of data, like our more common normal distribution. For example, if we want to count the per capita income of a certain city, in fact, most of the per capita income It is within a certain range, and only a small part is at the lowest and highest. In fact, this is the meaning of the median.

When doing data analysis, if the difference between each data is small, the average value will be better representative; and if the difference between the data is large, especially if there are individual In the case of extreme values, the median or mode is better representative.

Maximum (small) value

The maximum (small) value is more common when doing data analysis work, but we don’t pay special attention to it. The best values ​​are analyzed as typical representatives and outliers, such as sales champions in the sales team, popular e-commerce products, etc.

11. Comprehensive evaluation analysis method

Convert multiple indicators into an indicator that can reflect the comprehensive situation for evaluation, such as enterprise economic benefit evaluation. Including principal component analysis method, data envelopment analysis method, fuzzy evaluation method, etc.

The above is the detailed content of Commonly used data analysis methods. For more information, please follow other related articles on the PHP Chinese website!