# Google Analytics - 1 Foundations - Module 4 ## 0. Overview ### Learning Objectives ### Glossary of Terms [[Business Task]] [[Fairness]] [[Oversampling]] [[Self-Reporting]] ## 1. Data analyst job opportunities ### Video: _Let's get down to business_ (0:49) This is an overview video of the upcoming course. ### Video: _The job of a data analyst_ (4:00) Video covers an overview of business industries that use data analytics. #### Coca-Cola Data mining from the new computerized vending machines. #### Google Of course. Google search. #### Small businesses City zoo example. The zoo's main problem was that on bad weather days, they found themselves overstaffed. On really great weather days, they could be understaffed. #### Healthcare city hospitals can use data to determine staffing etc. ### Video: _Joey: Path to becoming a data analyst_ (2:00) Joey had come to the REWS team at Google. I really enjoyed this video because I am starting to learn being a data analyst does not mean that there is one specific path that brings practitioners into the profession. ### Self-Reflection: _Business use of data_ In this exercise I was asked to reflect on a business that I interact with and to consider some customer experiences that could be improved with the use of some data analytics. ### Video: _Tony: Supporting careers in data analytics_ (2:00) Tony gave a nice talk about the motivations that drive him to develop young professionals ### Interactive activity: _Test your knowledge on data analyst roles_ This activity was a matching game, matching the data analytics problem with the industry it came from. <hr> ## 2. The Importance of fair business decisions ### Video: _The power of data in business_ (4:00) #### Question vs Problem! >[! cue] Issue vs Questions vs Problem >[! info] Issue >A topic or subject to investigate >[! info] Question >Designed to discover information >[! info] Problem >An obstacle or complication that needs to be worked out. These kinds of questions and problems become the foundation for all kinds of business tasks to solve as a data analyst. #### Business Tasks >[! cue] Def [[Business Task]] >[! info] Business Task > The question or problem data analysis answers for a business. Keep in mind that the *question* and *problem* relate to the terminology above. #### Example, City Zoo In this example, we know the problem already that unpredictable weather made it hard to forecast staffing needs. >[! question]- What would a business task for the City Zoo be for this problem? >>[! quote] Analyze weather data from the last decade to identify predictable patterns. >[! cue] [[Data-driven decision-making]] Now that the analysts have defined the problem and a business task around that problem, the analysts can apply the [[data analysis process]] to step towards solving the problem using data to enable leaders to make [[Data-driven decision-making|data-driven decision]] ### Video: _Rachel: Data detectives_ (2:00) Interesting note here that the analyst being interviewed gave the advice that; - when working with new data, can be overwhelming - if an angle doesn't work out, don't give up, try other approaches. ### Video: _Understand data and fairness_(5:00) A key responsibility is to ensure that analysis is fair. >[!cue] Def: [[Fairness]] >[!info] Fairness > Ensuring that your analysis does not create or reinforce bias. Why it is hard. There are no standard definitions of it. >[! question] Sometimes conclusions based on data is true but also unfair. What to do then? an ethical data analyst could look beyond what simply the data show. >[! cue] Critical Responsibility: think about fairness from the begining >[! Warning] Critical to ensure analysis is fair by looking beyond the top level conclusion and consider other contexts that could bias the final findings. So the advice here is to consider fairness as early as possible. #### Example of a good bias consideration >[! cue] example of Oversampling A research team at Harvard conducted a study that had the potential to be very biased. To overcome this they partnered with social scientists to understand how bias could be introduced into the research. Additionally, to ensure that the groups were represented fairly, oversampled non-dominant groups. This had the effect that all groups were represented fairly in the dataset ### Reading: _Consider fairness_ the reading presents some best practices to enable more fair analyses. The practices are presented in a table, but since these are notes it is not the scope to reproduce all the data. I will attempt to summarize each BP #### Consider all of the available data. >[! cue] Consider all data Part of the job is consider what data will be useful for the analysis. Excluding data because it doesn't seem relevant could introduce bias and unfairness. - The example case discussed a data team measuring traffic on holidays. The team realized that they were excluding weather. So, by including weather they had a more complete analysis which is more fair. >[! warning] Consider all the data #### Identify surrounding factors >[! cue] [[context]] is important Similar to the considering all the data, also consider all surrounding factors that could influence the insights you're gaining. - The example case was in regard to a bank analyzing staffing needs and using the bank holiday schedule as a key metric. Obviously only using the bank holiday would introduce bias in that they would be ignoring other cultural holidays that the population celebrates. This would mean that the analysis didn't represent fairly the population being studied. #### Include self-reported data. >[! cue] Def: [[Self-Reporting]] >[! info] Self-reporting >A data collection technique where participants provide information about themselves. Using self-reporting techniques can avoid [[Observer bias]] Also provides context to conclusions - The example application tells the story of a data analyst at a retail store that created a survey that included self-reported information from the participants. This avoided introducing conscious or unconscious bias into the study. #### Use [[Oversampling]] effectively >[! cue] Oversampling >[! info] Oversampling >The process of increasing the sample size of nondominant population This can be a great way to represent all segments of a group in the population. - The example case was of a fitness company oversampling the elderly customers in their study to ensure that the analysis would be fair for all age groups . #### Think about [[Fairness]] from the beginning. Data collection, cleaning, processing, and analysis all done with fairness in mind. ### Self-Reflection: _Case opinion_ 3 case studies were presented. The requirement is to thoughtfully respond to the prompts with a minimum of 40 words for each prompt. Because this is graded, my responses are private. Essentially, the course is asking us to think critically about a set of scenarios to identify the fair or unfair data collection that is being presented. ### Video: _Alex: Fair and ethical data decisions_ (3:00) Alex works on a data ethics team at Google. Data ethics is more than minimizing harm, its about beneficence. >[! quote] How do we actually improve the lives of people by using data? >[! question] Are the people represented in this data going to benefit? > Need to keep in mind that data is coming from people and thus...data are people. As analysts we have a responsibility to those people in those data. Additionally, need to think about how to protect other people's data private. Like actual ways that consent occurs, and actionable ways that data can be revoked. ### Quiz: _Test your knowledge on making fair business decisions_ 4 questions, untimed, 75% minimum required. #### Result: 100% Be careful when considering how to be fair in data analysis. <hr> ## 3. Explore your next job ### Video: _Data analysts in different industries_ (5:00) Important considerations in job searches. - industry - each industry has their own data needs. - Need analysts with different skills - determines the types of tasks you will be working on. - Think about your interests early in a job search. - Location - does the preferred industry have opportunities locally? - tools - travel - culture ### Reading: _Data analyst roles and job descriptions_ This is so that I can understand differing types of analyst within an enterprise. - Business Analyst analyzes data to help businesses improve processes, products or services. - Data analytics consultant - analyses the systems and models for using data - data engineer: prepares and integrates data from different sources for analytical use - data scientist - Operations analyst - you will want to .. ahem .. analyze job descriptions in the industry of interest. ### Video: _Samah: Interview best practices_ (2:00) Samah is a recruiter and gives some good advice. Prepare questions for the interviewer ahead of time. Reach out to the hiring manager or a recruiter. ### Reading: _Beyond the numbers: A Data analyst's journey_ Actually not a reading. We are asked to view a TEDx talk about a career journey. <iframe width="560" height="315" src="https://www.youtube.com/embed/t2oOFs4WgI0?si=6k3RYpkVjlqubraK" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe> <hr> ## 4. Module 4 Challenge ### Glossary of Terms [[Business Task]] [[Fairness]] [[Oversampling]] [[Self-Reporting]] ### Module 4 challenge 8 Questions, 40 minutes timed. #### Result 100% ## 5. Course Wrap-up ### Reading: _Join the community_ Coda and recap Learning about the key tools and processes that form the core of data analysis. >[! cue] Community [Link](http://www.coursera.support/s/group-invite?id=MEY5VkgwMDAwMDAwMWRkMEFB) Advised to join the community. Would be beneficial if job leads and the like were available there. >[! cue] Study group [Link](https://lu.ma/GoogleCareerCertificatesStudyGroup) Also urged to join study group. If anything this will help keep me on track. The study groups are offered organized on Luma and conducted on Google Meet. Each session is 2 hours long, broken into two 50 minute sessions. Sample agenda below. >[! warning] Sample study group Agenda >- Intro and Weekly Focus (6m) >- Sprint 1 (50m) mics off camera on (optional) >- break (10m) >- Sprint 2 (50m) >- Wrap up (4m) Attempt regular attendence. ### Reading : Course 1 glossary [[Nexus - Google Analytics Certificate Course#1.0 Glossary of terms from this course|Course 1 Glossary]] ### Video: _Congratulations!_ Tony gives a great send-off! A lot of data analysts try to make things perfect the first time. They make assumptions , better to be humble and inquisitive and to ask questions. The team is the best resource. Nervous about he would be perceived by asking questions. There are no bad questions. ### Reading: _Coming up next ..._ ![[course 1 complete.png]] <hr> >[!summary-top] Module 4 key take aways >1. Fairness and completeness are incredibly important. >2. [[Business Task]] is a new concept for me. Remember the distinction between, Problem, Question, and Issue: see [[#Question vs Problem!]] >3. Asking for help from Team should be a priority over going lone wolf. Don't make mistakes.