Image through an ideogram
# Entry
When you hear the word, you probably think about two words: programming and statistics. In fact, the condition of learning statistics of learning often discourages people from conducting a data career. This does not facilitate that most of the descriptions of scientific positions make you think you need a doctorate in statistics to develop in this role when reality is completely different.
At most scientific positions, especially in technology companies, focusing on product development, you need to know Used statistics. This includes the employ of existing statistical framework for solving business problems. This is different from academic statistics (by hand calculating intricate formulas). Instead, you just have to understand what the concept means, how to calculate it with existing libraries and how to interpret it. Here is an example: in the most practical scenarios of data sciences, it is enough to understand what a P value of 0.03 means and how to employ it to make a business decision, and not necessarily know how to calculate it manually.
In this article I will give examples of how I employ statistics in my work in learning data, along with the resources that I used to obtain this knowledge.
# How do I employ statistics in my work. Data learning
// Experimentation
Most technology companies (Google, Met “have a gigantic culture of experiments. They test strictly before making functions changes.
When conducting A/B tests, I need to know statistical concepts such as:
- Statistical power to determine the size of the sample required for the experiment
- Levels of significance, PI values trust in the field of decision making
There are times when P values may not tell a full story in which you will have to learn more intricate forms of analysis, such as differences in differences (DID). However, these are the concepts that I chose at work, reading articles, asking questions and discussions with older colleagues. You can’t learn and remember about every concept required by courses and even a university degree. I suggest you take basic concepts that are required to interview data learning and learn the rest at work.
// Modeling
Building machine learning models requires knowledge about statistics. However, in my experience, it is enough to have practical knowledge about machine learning models, instead of learning the theory of these algorithms and the way they are created.
Of course, this does not apply to every industry. A scientist of data working in a specialist sector, such as forecasting, biostatism or econometrics, must have deep statistical knowledge about their field.
In my experience, during work in products or technological companies, the business impact and interpretation of these models were focused more than on mathematical exacting.
// Data analysis
I also spend a significant amount of time analyzing data to understand how users interact with the product, providing recommendations on how to improve this experience. This usually includes descriptive statistics in which I create visualizations, I segment customer segmentation and compare data distribution. Most of the questions related to data, such as “why customer detention has dropped in the last 3 months”, can be solved with straightforward visualizations and does not require the employ of sophisticated statistical methods.
# Three resources to learn statistics on data education
I have a degree of computer science and I have not learned much statistics. All my statistical knowledge comes from the resources that I found online, and I have developed a list of the most helpful:
- Introduction of Udicita to statistics It is recommended for complete beginners and includes descriptive statistics, application statistics and probability
- Status It is helpful when you want to learn specific concepts. For example, if you want to find out how regression works, you can find 20-minute tutorials that are specific to the topic on this channel
- Statistical learning at EDX This is another great course that you can control for free. This learning path teaches the employ of statistical concepts in Python, which makes it vital for most scientific tasks
# To go
While the idea of learning statistics on data learning may seem intimidating, most tasks on data require known statistics used, i.e. the possibility of using statistical concepts to solve business problems. In my experience, this knowledge can be easily acquired via online courses and does not require a master’s degree in statistics.
The resources listed in this article should be enough to go through some of the scientific interviews of statistics. Each knowledge can also be acquired at work, constantly reading articles and articles on this subject, working with existing frames in your organization and learning from older data scientists.
Natassha Selvaraj He is a scientist of self -taught with passion for writing. Natassha writes about everything related to data, a true master of all data topics. You can connect with it LinkedIn or check it YouTube channel.