# # Google Analytics - 2 Ask - Module 1 - Ask Effective Questions ## 0. Overview >[! quote] Data analysts are constantly asking questions in order to find solutions and identify business potential. In this part of the course, you’ll learn about effective questioning techniques that will help guide your analysis. ### Learning Objectives >[! info] Module 1 Learning Objectives >- [ ] Explain the characteristics of effective questions with reference to the [[SMART methodology]] >- [ ] Discuss the common types of problems addresed by the data analyst >- [ ] Explain how each step of the problem-solving roadmap contributes to common analysis scenarios >- [ ] Explain the [[data analysis process]], making specific references to each phase; [[data analysis process - Ask]], [[data analysis process - Prepare]], [[data analysis process - Process]], [[data analysis process - Analyze]], [[data analysis process - Share]], [[data analysis process - Act]] >- [ ] Describe the key ideas associated with [[Structured thinking]] including the [[Problem domain]], [[Scope of work (SOW)]], and [[context]] --- ### Glossary for Module 1 - [[SMART methodology - Action-oriented|action-oriented question]] - [[cloud]] - [[data analysis process]] | [[data life cycle]] - [[Leading question]] - [[SMART methodology - Measurable|measurable question]] - [[Problem types]] - [[SMART methodology - Relevant|relevant question]] - [[SMART methodology]] | [[SMART methodology - Specific|specific question]] | [[Structured thinking]] - [[SMART methodology - Time-bound|time-bound question]] - [[Unfair question]] <hr> ## Problem-solving and effective questions ### Video: _Introduction_ Introduction to instructor Ximena, Google Finance data analyst and to course. We will be learning. - the difference between effective and ineffective questions - effective questions will enable you to maximize each step of the [[data analysis process]] #### Refresher Ask. in the ask step we define the problem that we are solving and make sure that we fully understand [[stakeholder]] expectations --> which will help us stay focused on the problem we are solving. The rest of the video Ximena discusses other things that will be discussed. [[Structured thinking]], [[Spreadsheet]], and more work in [[SQL]] >[! cue] [[Structured thinking]] preview Finally, >[! quote] Structured thinking is the process of recognizing the current problem or situation, organizing avaialbe information, revealing gaps and opportunities, and identifying the options. In this process ==you address vaugue, complex prblem by breaking it down into smaller steps which lead you to a logical conclusion.== ### Reading: _Course 2 Overview: Set your expectations_ - this reading sets up an overview of the information to come, the types of activities used to present the material, and establishes best practices for completing the course. - Because each step is documented in a [[learning in public]] style, reiteration is not needed here. - The material in the overviews is captured in the parent note [[Nexus - Google Analytics Certificate Course|Google Analytics Certificate]] under the second course headings. ### Interactive Activity: _Refresher: Certificate Roadmap_ It is this document that informs my self-rating topics, not the individual topics. ### Reading: _Evaluate your current data analytics skills_ This is another evaluation on current skills. In fact it is the same one from the first course. No notes taken here. For anyone else reading, there are 11 questions and if you affirm each of them confidently, then you would be better off in the Advanced analytics certificate and another in the Business Intelligence Certificate program too. <hr> ## Take action with data ### Video: _Data in action (5:00)_ - case study. - How problem solving relates to each phase of the [[data analysis process]] #### Ask - begin by defining the problem - zoom out to gain context - Enables analyst to ==focus on the real problem and not just the symptoms.== - Then identify and work with stakeholders to clearly define the problem #### Prepare - needed more information about the target audience. (more context this time on the audience) - after collecting data about the target audience, then collected data on different advertising methods #### Process In this phase of the analysis process, the analyst then cleaned the data to ensure that no inconsistencies etc were going to get in the way and removing any outliers that might make the final result unfair. #### Analyze The analyst then used her analytical skills to to confirm the target audience and then determine an answer to the problem. ==knowing the context, and stakeholder expectations== further enabled her to make the right recommendation that was appropriate for the company. #### Share. To share, the analyst created compelling visuals to explain the recommendation that was being given. The clear and compelling visuals enabled the [[stakeholder]] to understand the solution. #### Act Now it was time to act, the company then implemented the change which led to real results. ### Reading: _From issue to action: the six [[data analysis process]] phases_ More detail provided on applying the 6 phases of the analysis process! #### Ask Considerations: - define the problem - ensure full understanding of [[stakeholder]] expectations - Focus on actual problem - a way to do this is to zoom out - look bottom up and top-down points of view - Collaborate with stakeholders and keep line of communication open. - Consider fairness here, ask about higher contexts that might make the analysis unfair Questions to help: - What are my stakeholders saying their problems are? - Note that this question is prompting what the _stakeholders_ think not what **you** think. - Now that I have identified the problem/issue/question, how can I help the stakeholders resolve the question? - This is a question of partnership and of service to the stakeholders. How can I help, not how can I save. helping is a form of collaboration, no one wants to be dictated to. #### Prepare The analyst decides the data needed to be collected to answer the question. The analyst should also understand how to organize the data so that it is useful. Considerations: - What metrics to measure - where is my data, do i need permission to access it? - How can I protect my data when I have it? - What is the plan to retain the data, how will I destroy it? - Is there sensitive information the data? - Is there protected information in the data? - Consider fairness here. - Is the data showing the complete picture are is there more that needs to be collected/accessed or created? Questions to help: - What do I need to figure out how to solve this problem? - What research do I need to do? - Is my research including or fairly representing the data that will be processed? - Am I at risk for introducing bias? How would I address bias? #### Process In the process phase, the analyst will transform the data into a usable form. This is where we conduct data cleaning and remove outliers and explain why nulls exist. Considerations: - using spreadsheet functions to find incorrectly labeled data. - Using SQL functions to check for spaces. - Duplicated entries. - Considering fairness: Does bias exist? check, check, then check again. Questions to help: - What data errors or inaccuracies might get in the way of getting the best possible answer to the problem I am trying to solve? - How can I clean my data so the information i have is consistent? #### Analyze In this phase, we want to think analytically about the data. Considerations: - sort/format data - combine data from multiple sources - create tables with the results. Questions to help guide: - What story is the data telling me? - How will my data help me solve this problem - Who needs my company product / service What type of person is most likely to use it? #### Share This is all about communicating the results of the analysis in a compelling format that fits with the audience. Questions that guide: - what kinds of visuals would be most impactful to my stakeholders considering this problem? - How can I make this easy to understand and engaging? - What would I need if I were a listener? - Remember GPS/ and the curse of expertise? #### Act This could mean making a recommendation based on the data. Question to consider: - How can I use the feedback I recieved during the share phase to acutally meet the stakeholder's needs and expectations? >[!cue] Def [[Structured thinking]] >[! info] Structured thinking >breaking the a larger process down into manageable steps >There are four basic activities in this process: >1. Recognize the current problem >2. Organize available information >3. revealing gaps and opportunities >4. Identifying your options > >The data analysis process is form of structured thinking. ### Video: _Nikki: The data process works_ (2:00) The presenter discusses the [[data analysis process]] by giving a case study that they actually worked on at Google. ### Quiz: _Test your knowledge on taking action with data_ untimed quiz, 75% minimum score, 4 questions. #### Results: 75% on this quiz or 3 out of 4. I missed a question about _reaching a target audience_. The reason I missed it is because I thought the question was asking a different question. The question described a scenario the plainly explains reaching a target audience. My mistake was thinking in broader terms of what is the analyst helping the company do (data driven decision making) I understand that the correct answer was "reaching a target audience". <hr> ## Solve problems with data ### Video: _Common problem types_ (5:00) Understand the kind of problem Focus on 6 common types >[! cue] 6 [[Problem types]] 1. [[Problem types - making predictions|making predictions]] 2. categorizing things 3. spotting somehing unusual 4. identifying themes 5. discovering connections 6. finding patterns ### Reading: _Six common problem types_ ==Analytics is so much more than plug-n-chug== First step is most important. **Understand the problem** From there use problem solving approach: decide what data needs to be included, how data can be transformed, and how the data will be used. >[! cue]- 6 problem types >- [[Problem types - making predictions|making predictions]] >- [[Problem types - categorizing|categorizing things]] >- [[Problem types - spotting something unusual|spotting something unusual]] >- [[Problem types - identifying themes|identifying themes]] >- [[Problem types - discovering connections|discovering connections]] >- [[Problem types - finding patterns|finding patterns]] #### Making predictions Example problem statement: > What is the best advertising method to bring in new customers? Data ideas: - location - type of media - prior advertisement performance - market trend #### Categorizing things Example problem statement: > How can we improve customer satisfaction? Data ideas: - customer service call satisfaction score data - customer data - self-reported survey data, keyword mining Could use additional techniques to identify clusters / patterns from processed dataset. #### Spotting something unusual Example problem statement: > How can we be more aware of human trafficking? > How can we avoid future instances of fraudulent activity? >[! question] a tougher problem statement to assess because how do we know when to look and what to look for. The anomaly detection algorithm would be what the analyst would solve? which algorithm to use? Data ideas: - historical data - fraudulent activity oversampling to balance dataset #### Identifying themes Example problem statement: > What features should be prioritized on _projectX_ ? >[! tip] Theme identification is dependent on categorization. Taking the categories and grouping into broader concepts. Data ideas: - Customer survey data on feature requests - keyword mining. #### Discovering connections This example is a little different: - an online retail company that works with a third party logistics vendor wants to improve on time delivery to their customers. Example problem statement: > how to improve ontime delivery metric for customers Data ideas: - weather - staffing - shipping - get third party shipper data? - wait times at particular hubs (first party data) Approach: because there is an opaque part of the problem, the third party vendor data, we want to find the connection that is likely to lead to delays. #### Finding patterns Example problem statement: > How can we improve our trade execution for our customers? Data ideas: - Customer trading records - Customer channel data - customer satisfaction (CSAT) - employee staffing data Approach: - try and discover if any patterns exists with poor execution trades vs cleared execution, Channels used, timing, etc. Conduct gap and root analysis. output dashboard, outcome, improve kpi of trade execution. ### Video: _Continue exploring business applications_ (6:00) Recap of what was taught so far. Problem types. [[Problem types - finding patterns]], [[Problem types - discovering connections]], [[Problem types - making predictions]], [[Problem types - categorizing]], [[Problem types - identifying themes]], [[Problem types - spotting something unusual]] This is a review of the reading from the last section. ### Interactive activity: _Name the problem type_ In this activity, I am to identify the six problem types. These are presented as flashcards. 1. making predictions 2. categorizing 3. spotting something unusual 4. identifying themes 5. discovering connections :: identifying similar challenges across different entities. 6. Finding patterns. ### Video: _Anmol: From hypothesis to outcome_ (2:00) This video's presenter talks about hypotheis testing/ AB testing from a high level The point is that the team uses data to discover what the customer needs first, proving what needs are true and what is not needed, then using the insights to drive positive outcomes (marketing teams producing content that is user driven) ### Quiz: _Test your knowledge on solving problems with data_ 4 questions, timed. #### Results 100% - Finding Patterns depends on historical data while making predictions is using data to make informed decisions about the future. - [[Problem types - finding patterns]] - [[Problem types - making predictions]] ## Craft effective questions ### Video: _SMART questions_ Data analysts ask lots of questions. The more questions you learn about your data. Some are more effective than others... Examples of ineffective or not effective questions [[Leading question]], questions that lead you to answer in a certain way, but limits you to discovering anything useful in the answer. [[Closed-ended question]], a question that is answerable with only a "yes" or "no" response meaning you don't learn very much by getting an answer. Know the difference between effective and ineffective questions is very important. Effective questions follow the [[SMART methodology]] SMART is an acronym S = Specific M = measurable A = Action-oriented R = Relevant T = Time-bound [[SMART methodology - Specific|Specific]] questions are simple, significant and focused on a single topic or a few closely related ideas. - If the question is too general, try to narrow down and focus on one element. [[SMART methodology - Measurable|Measurable]] questions can be quantified and assessed. - Example of ineffective: "Why did a recent video go viral?" - Example of more effective question: "How many times was our video shared on social media in the first week it was posted?" [[SMART methodology - Action-oriented|Action-oriented]] questions encourage change. - Example ineffective: How can we get customers to recycle our product packaging? - Example effective: What design features will make our packaging easier to recycle? [[SMART methodology - Relevant|Relevant]] questions matter, are important, and have significance to the problem you are trying to solve. - ineffective example: "Why does it matter Pine Barrens tree frogs started disappearing" - irrelevant because the answer does not get us close to our solution state. - effective example: What environmental factors changed in Durham, NC between 1983 and 2004 that could cause Pine Barrens tree frogs to disappear fro the Sandhills regions? [[SMART methodology - Time-bound|Time-bound]] questions specify the time period to study. - the prior example of effective questions designated the time period of 1983 to 2004. Knowing this will limit our data set to the appropriate years >[! question] >Is the SMART methodology presented here the same as the one coined by Peter Drucker? >[! warning] We need to also consider [[Fairness]] #### Examples of unfair questions - [[Leading question]], makes it difficult to answer if there is disagreement. Fair questions also means that the language is clear and accessible to others. Don't muddy the waters by using jargon. Make the questions easy to understand. ### Reading: _More about SMART questions_ cases and example uses of the [[SMART methodology]] all in one go. My problem statements should be SMART. #### Example 1. **What features do people look for when buying a new car?** Run through each of the SMART criteria and see if I can do better. - Specific - Measurable - Action-oriented - Relevant - Time-bound So, the problem statement should not be all encompassing, especially if that makes the main questions confusing. ### Practice Quiz: _Self-reflection: Practice working SMART_ In this activity I am presented with a scenario in which to apply what I have recently learned. #### Results: These activities are not graded, but are pass fail. My result is Pass. #### Additional info: After submission there is additional information provided as a coda. To kick start your questions, some suggested topics were given; - Objectives - Audience - Time - Resources - Security considering asking questions on the above topics effectively to get an even better starting place from a vague starting point. ### Video: _Evan: Data Opens Doors_ 3 diff core roles - data analyst: sql spreadsheet databases - data engineer, turn raw data into actionable pipelines - data scientists: model finished data ### Practice Quiz: _Self-reflection: Ask your own SMART questions_ This was another self-reflection activity which is technically pass/fail. For this activity I was asked to contact a data professional to ask SMART questions, The purpose of this activity was to practice having and thinking about data-driven conversations. ### Practice Quiz: _Test your knowledge on crafting effective questions_ 4 questions. timed. #### Results: My result is 100% ## Module 1 Challenge ### Glossary for module 1 - [[SMART methodology - Action-oriented|action-oriented question]] - [[cloud]] - [[data analysis process]] | [[data life cycle]] - [[Leading question]] - [[SMART methodology - Measurable|measurable question]] - [[Problem types]] - [[SMART methodology - Relevant|relevant question]] - [[SMART methodology]] | [[SMART methodology - Specific|specific question]] | [[Structured thinking]] - [[SMART methodology - Time-bound|time-bound question]] - [[Unfair question]] ### Quiz: _Module 1 Challenge_ 80% required. Timed 40minutes, 10 questions #### Results 100% <hr> ## Summary and takeaways >[! summary-top] Module 1 Key Take Aways >1. There are 6 common problem types: [[Problem types - finding patterns|Finding Patterns]], [[Problem types - making predictions|Making Predictions]], [[Problem types - discovering connections|Discovering Connections]], [[Problem types - identifying themes|Identifying Themes]], [[Problem types - spotting something unusual|Spotting something unusual]], [[Problem types - categorizing|Categorizing]] >2. Watchout for the subtelty on [[Problem types - making predictions]] vs [[Problem types - finding patterns]] and [[Problem types - categorizing]] vs [[Problem types - identifying themes]] >3. Be [[SMART methodology|SMART]] and ask effective questions be ensuring that questions are [[SMART methodology - Specific|specific]], [[SMART methodology - Measurable|measurable]], [[SMART methodology - Action-oriented|action-oriented]], [[SMART methodology - Relevant|relevant]], and [[SMART methodology - Time-bound|time-bound]] >4. Always consider [[Fairness]] when thinking of questions and avoid [[Leading question]] or [[Closed-ended question]]