Ten Rules for Thinking Differently About Numbers

💡 The book contains high-quality content and many ideas correspond to Google’s data analysis course, making it particularly impactful to read.

How to Make the World Add Up: Ten Rules for Thinking Differently About Numbers

Opening the Door and Seeing the Mountain: Ten Rules

• First, we should learn to stop and pay attention to our own reactions to certain statements, rather than simply accepting or rejecting them based on our feelings. This allows us to be more open-minded and consider different perspectives.
• Second, we should find ways to combine a “bird’s-eye view” and a “worm’s-eye view” when looking at things. This means applying a statistical perspective to understand the broader picture, while also incorporating personal experience to understand the nuances and details of a situation.
• Listen (listen in all directions, focus on listening to the other person, do not listen and think about what to say next; ask questions to clarify the problem).
• Third, when obtaining data, we should look at the labels and ask ourselves whether we understand what these data are really trying to portray.
• Fourth, we should seek comparison and context, and any argument must be seen from a broader perspective.
• Fifth, we should look at the background of the statistical data and understand how these data came about – and which data may have disappeared.
• Sixth, when data is placed in front of us, we should ask, who is missing from this data? If we include them, would our conclusion be different?
• Seventh, we should be wary of the sharp problems with the methods and drivers of large data sets. We need to understand that if they are not open and transparent, we cannot trust these algorithms and data.
• Eighth, we should pay more attention to the cornerstone of official statistical data – and the statisticians who bravely defend these data.
• Ninth, we should recognize that the numbers we are using are not always accurate, and we need to consider their uncertainty.
• Tenth, we should avoid using only one number to prove something, and we need to consider the whole picture.

Above all, curiosity is needed to drive it

Curiosity occurs, usually in the gap between the known and the want to know, which is the sweet spot of curiosity; using this point can make us more interested in everything.

Content summary and impressions

Plain and simple theory

We easily regard our own viewpoints as universal viewpoints, and this subjective idea often conflicts with “reality”, so no matter when predicting or interpreting data, we need to maintain a more objective heart; of course, combining “worm’s-eye view” is not bad, but don’t forget the “bird’s-eye view”.

Emotion

When we see a certain data, it is likely that a certain emotion will be caused by our own position, whether it is complacent or defending, which will make us have different interpretations of data, so when facing data, we need to calm down a little and not make conclusions too quickly.

Expert traps

When we have a certain professional knowledge and are full of motivation, it is likely to lead to worse bias; constantly check whether our position comes first or data comes first.

What is the real problem

When we want to answer a question, we must understand the true meaning of the question.

Definition of data text

When we are examining data, we need to pay attention to whether the definitions of words in the data are the same as our imagination, otherwise the data analysis is likely to be empty.

Data landmarks

When we face an unfamiliar data, we need to find references to know whether the number is large or small.

Source of data

If the source of data is biased, the result will also be biased; if the data comes from others, we need to examine their conditions, and if it comes from us, we also need to find potential assumptions as much as possible.

Lack of Big Data

When we see a lot of data, we assume that it can represent the whole, but we must always think about what is missing in it.

The importance of data correlation

Even if we find a trend between the data, we do not understand the reason, we may convince ourselves that this is the case, but the relationship between them is still important, and we cannot simply assume their relationship, but we need to understand the reasons.

Big Data Misconception

As it becomes easier to obtain data, we may mistakenly believe that data is more useful. However, no matter how large the data is, there will still be small data problems, and the situation may be worse.

Importance of Public Data

If data is only for government use, it is useless for improving the government. When this data (assuming it is reliable) can be provided to the people, we can supervise the government well. In fact, this way the government can also improve itself more efficiently.