How to select the appropriate statistical chart in the website Analysis report

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

Because of the diversity of data collected by different web analytics tools, statistical charts based on these data generation are also different, and when web analysts make various reports, they often struggle to choose the right charts to express their data demands, so we need to understand some of the most common web analytics (not just web analytics, Data analysis in other areas also applies to the selection of these charts. The features, usage and points of attention of statistical charts. The following charts are used primarily in data analysis.

Types of charts

Line chart: Using when analyzing trends in data by time series

Column chart: Specifying an analysis axis for data size comparisons using the

Pie chart: Specifying an analysis axis for proportional comparisons

Instrument Chart: Use when focusing on performance performance of an indicator individually

Line chart

The line chart is mainly used in the case of analyzing the trend of index value according to the time series. Typically, the x-axis is set to time, and the y-axis is set to other metric values. Analyze the number of page views, visitor number, conversion (rate) and other indicators of the overall trend of change, more use line chart. These index values are expressed by line chart, you can clear the trend of each hour, day, week, month or year, to get similar "weekdays more visits, weekend visits less", "this month's conversion is lower than last month nearly 10%" and other analytical conclusions.

So then let's go and see some of the line charts that the Web analytics uses (the x axis is set to time).

First, the y-axis is set as a line chart for the number of page views. After the line chart, the number of page views increases or decreases at a glance.

(※ Line chart)

The following figure shows the number of page views that cover both initial access and re access. In the same unit, these metrics can be combined to display in a single chart.

(※ Subdivision Line chart)

This makes it possible to refine and analyze the reasons for some trend changes. In the figure above, it can be found that the short peak of traffic on August 16 was mainly from the first-time visitors.

Next look at the picture below. Is "This August" and "last August" every day of page browsing number of data. It can be seen that this year not only the overall flow has been improved, but also effectively alleviate the weekly traffic reduction phenomenon.

(※ Contrast Line chart)

Considerations when doing line charts

Note the following when you make a line chart.

1. The indicators in the chart are to be expressed (that is, the values represented by the y-axis).

2. A line chart is recommended when there are more than 5 items on the x-axis, and can be used when there are less than 5 items.

3. In a set of line charts, if the polyline exceeds 4, there is a duplication between the lines that will not be seen clearly. In this case, you can split into two sets of line charts to perform.

Column Chart

The column chart is mainly used in the case of a specific axis to compare the size of the metric value. Column chart is one of the most commonly used charts in Web analytics. Column graphs can be divided into vertical column, horizontal column and cumulative vertical column. Here are some of the characteristics of these column graphs.

The first is "horizontal column diagram". The horizontal column chart is also called the bar chart, is generally used to represent a kind of project horizontal contrast, for example by the visit quantity to the page rank, according to the conversion rate size to the advertisement media rank and so on. The x axis of a horizontal bar chart usually represents a scale ruler that determines the size of a value. The following figure is the ranking of the site pages by the size of the traffic:

(※ Horizontal Column chart)

The second is "vertical column diagram". Vertical columns and line charts are similar in usage and are commonly used to denote the change in the indicator values of time series. The difference is that if you have less time on the x-axis (for example, less than 5), you can choose to use a vertical bar chart, or choose from the nature of the data and the focus that the chart wants to express: the vertical column is biased toward the number of representations, and the line chart favors performance trends.

As you can see here, the "vertical column" represents the relationship of data over time, and the "Horizontal bar chart" indicates that the item is compared to the data at a certain point in time or time period. Some people might ask, "Does the data in the horizontal bar chart show up in a vertical column?", theoretically, but usually the name of the item in the horizontal bar chart (for example, the page title in the previous example) is long, and if it is difficult to show through a vertical bar chart, the behavior of the row may occur.

The following figure is a vertical column that shows the number of accesses over time.

(※ Vertical Column chart)

There is also a column diagram called Cumulative vertical column. It is convenient to use the Cumulative Vertical column chart when we want to confirm the details of a particular value. The following figure is a breakdown of the number of visits by site (advertising entry, direct access, search engine entry), which makes it easy to judge the details of how the site enters.

(※ Cumulative Vertical Column chart)

Considerations for making column diagrams

Please note the following when making the column chart.

1. Pay attention to the distinction between vertical and horizontal column graphs

2. The number of columns caused the gap between the columns is very small, you can choose to use line chart

3. "Quantity" and "proportion" are best made into different column charts.

Pie chart

Pie charts are mainly used when analyzing the proportions of components in the overall metric value. The feature is to be able to discern at a glance which item accounts for the largest proportion. It is easy to find the site's entry page, which is mainly the "Digital Forest Blog" page, with a ratio of nearly 40%.

(※ Pie chart)

Considerations when making pie charts

Please note the following when making a pie chart.

1. Because the pie chart is to confirm the proportions rather than the quantity, the pie chart represents the proportions, and if you want to represent the total value, you can add a row n=xxxx such a description.

2. The constituent elements in a pie chart are sorted in descending order of percentages (except for "other")

3. Because people are used to looking at things clockwise, it is best to place the most important content near the 12 point

4. When there are more than 10 constituent elements in the pie chart, the data in the back row is summarized in the "other" notation.

Instrument Chart

The instrument chart is used to focus on the performance of a single indicator. The characteristic is that it can give the safety range and the warning range of the index.

For example, the performance of the bounce rate is divided into three performance regions by 25%, 50%, and the three performance areas are differentiated by green, yellow and red respectively, where red indicates the alert area. From the following figure can see the September jump rate has more than 50% into the alert area, need to pay attention to.

(※ Instrument Chart)

Considerations when making a dashboard chart

Please note the following points when making the instrument chart.

1. Attention to the Division of performance areas, generally using two edge values into three regions

2. The Red alert area appears on the left or right and is related to the target's business attributes, such as the higher the bounce rate, the closer to the alert area, so the red area is on the right.

Finally, citing a passage from McKinsey's "Talking in a chart", we summarize the selection of the chart:

(1) A chart is a form of language that exists to express what you want to express more quickly and better than a table.

(2) Decide the chart is not the data nor size, but you want to explain the topic

(3) The chart is not expensive, only use when the chart can help you to express the theme

(4) The chart is an intuitive teaching aid, but it does not replace writing and narration, and it can play an important role in helping you communicate your theme.

Special Note: All of the above chart screenshots from the Web analytics tool-dimensional analysis dashboard.

(Copyright GUI Lin website Analysis blog All, welcome to reprint, but reprint please specify the source.) )

Original: Http://blog.digitalforest.cn/wangzhanfenxi-tongjitubiao