Section 003, Instructor: Ermira Zifla

Weekly Question #3: Complete by February 8, 2017

Leave your response as a comment on this post by the beginning of class on February 8, 2017. Remember, it only needs to be three or four sentences. For these weekly questions, I’m mainly interested in your opinions, not so much particular “facts” from the class!

If you sign in using your AccessNet ID and password you won’t have to fill in the name, email and captcha fields when you leave your comment.

In your opinion, what was the most important takeaway from last week’s discussion on bias and the Signal Problem? Give a recent example of how you’ve used data to make a decision, and a hidden (or not so hidden) bias that was likely present when making your decision. Did this result in a signal problem regarding your analysis and decision-making? Describe how you might have counteracted the bias and Signal Problem for your particular circumstance.

40 Responses to Weekly Question #3: Complete by February 8, 2017

  • I found that the most important thing in which I took away from this week’s discussion involving bias and Signal Problem was that we are unable to remove bias from our data sets, and though a larger data pool may reduce error, we are still unable to completely eliminate error. Also, in addition to that, there are variables in which will always go unaccounted for in any type of experiment, and while this too falls under the data error category, I find that the unaccounted variables, whether they be dependent o independent are very crucial in implementing diversity within a data pool. On a regular basis, I find myself making decision correlating with managing risk. I use previous experiences to calculate a situation’s outcome and I decide my actions and behavior according to the amount of risk taken and the reward as well. My use of previous experiences is a bias as if something ended badly in a previous situation similar to a current or future one, I would be less likely to engage in a possibly disastrous situation. Just the same as if a situation was good, I would be less likely to deviate from my previous actions in order to receive desired results. Now while this data doesn’t pertain to numerical information, it still is data as it is looked upon in reference in order to make a logical or ethical decision. Note, that while this data can be helpful, there could be an error in my decision making often as there is no way to predict the outcome or result o an independent variable. The only way to combat such decisions is o balance a little of what one knows and surveying the current situation and combining those two intuitions in order to select the best decision overall. Note that human intuition lacks the accuracy to that of numerical data.

  • The most important takeaway for me is the existence of the signal problem which i was not aware of. Especially the effect that this has on self reported data, which is very prevalent in aggregators like yelp and in political poll. Recently, I had to replace a computer part and used a review aggregate site to choose the part. A bias that I know I had is a brand bias and bias due to my small price range. I do not believe that these biases created a signal problem

  • Profile photo of Whitney Libebe

    My takeaway from last week’s discussion on bias and the Signal Problem is the importance in assessing and analyzing data, instead of just accepting it for what it displays. I was looking to buy a new skincare product last week and went on the company’s website to look at reviews. A blatantly obvious bias I quickly noticed was that the reviews were mostly rated either 5, or 1 stars, meaning the reviews had an extremely positive or an extremely negative experience. This was a huge signal problem regarding my decision making because I wasn’t sure whether I’d have a great experience with the product, or a completely horrible one. I tried to counteract the bias by averaging the numerical rating of the reviews. Because I was satisfied with the product, I ended up going to the store and trying a sample.

  • Profile photo of Lily Heinemann

    What I took away from last week’s discussion on the signal problem and bias was that we must take the necessary steps to try to eliminate bias from datasets and to try to account for any signal problems with data before they happen. Whenever I plan to buy something online, no matter what it is, I always try to read the reviews thoroughly in order to obtain the thoughts and opinions of other people who have purchased the product. The problem with many online reviews is that the people who submit them are either really pleased with the product they received or they’re really displeased. The majority of reviews are not based in the middle ground of grading (such as 2-4 stars,) but they’re either 1 star or 5 star reviews. It normally results in a slight signal problem because the people who leave the reviews have very strong feelings about the product, as opposed to those who were merely satisfied instead of amazed. I tend to counteract the signal problems with online product reviews by seeing how valid peoples’ reasons for reviewing are, and making a decision based on the less extreme reviews.

  • The most important takeaway from last weeks discussion on bias and the signal problem is that it is important to assess and analyze the data instead of just using whatever data pops up. An example of how I used data to make a decision was last week when I was deciding which restaurant to go to for “Restaurant Week.” I looked at all of the reviews for different places but the problem that may occur is the bias comments that were left because there are many different reasons why a person could of had a good or bad time. I feel the best solution is to go explore the restaurants and see for yourself, regardless of what is said online.

  • I found that the most important thing in which I took away from this week’s discussion involving bias and Signal Problem was that our technology gets to know us so well by the things we search for and how some data will never go against what we think because of this. An example of how I used data to make a decision was when I was looking at restaurants for Valentine’s Day dinner. I looked at places in Philly where the reviews are good and somewhere that was affordable. I was looking at the reviews and all the comments weren’t that helpful to me. Some said the food was good, some said the food was okay, but overpriced. It is hard looking at reviews because of bias. Everyone is different and everyone has their opinions.

  • In last week’s discussion on bias and signal data, the most important takeaway was that we must be aware of and actively seeking out hidden biases in data. In the past, when I’ve looked at data I always usually assume it is correct and represents the full picture. I now know that I have to look more deeply at the data to determine its accuracy and if there are any biases that are affecting the results. A recent example of when I’ve used data to make a decision is when I was researching reviews for different bus companies before buying a bus tickets to visit my friend in D.C. In the reviews, a bias that was present was personal experiences and their expectations. An example is that some people commented that the budget bus companies were awful because they had less leg space, but for a budget company this is expected, so the reviews were biased based on their personal expectations. A signal problem is that bus reviews are typically only going to be written by people who do not have their own car, so a large population is left out and the sample size is a lot smaller. I counteracted the biases by going ahead and buying a ticket for the budget bus company in order to save money. Since I know what to expect from budget companies, the reviews did not influence my decision.

  • Profile photo of James Dean

    The most important takeaway from last week’s lesson for me was how prevalent bias in data is. Sometimes a bias can be obvious, such as in the political and marketing fields. Other times though, a bias may be very hard to see or realize that it exists. There are also times when the bias is not mutually exclusive and several biases affect the data in question. Recently, when choosing an institution to attend to complete a BA in Finance, I had to compare data from a few different Universities. Although most of the data I reviewed favored Temple, I was heavily biased toward this University regardless. This was because my brother recently graduated from Temple last year. This resulted in a signal problem because I was well aware of the bias and tried to look past it when making my final decision. To counteract the bias and signal problem in my decision, I could have made my choice strictly based on ranking of the schools by a third party. This would show me which school was better, strictly on merit. Taking my bias toward Temple out of the decision.

  • Profile photo of Seth Keatley

    What struck me most about our discussion regarding bias in data was the notion that is an inherent part of data and information. These outlying pieces are in every dataset, because no data set is perfect. When I was shopping for Christmas presents this past December on Amazon, some of the reviews reflected a bias. The reviews on Amazon are sometimes dreadfully done, e.g. people review the fast shipping rather than the product, people not knowing how to use the product properly, people not actually buying the product, ect..I could have looked at the highest rated reviews to eliminate bias for the products because other users would have acknowledged that someone wrote a good review as opposed to someone who did not.

  • Profile photo of Walter Kirby

    What struck me the most in terms of bias and the Signal Problem was that there is no overall cure for bias. No matter what anyone does to make a sample better (i.e., have it be random, have a bigger sample, etc.) bias will always be present. There is no full proof method to fully eliminate bias. The best anyone could do is reduce it to the slightest extent but any sample has to be taken with a grain of salt because some bias will be evident in any set of data. And even if the bias is not glaring out at someone, there are little hints of bias that exist when a person thinks critically about the data they are analyzing. Recently, I was searching for a new pair of shoes and was researching reviews on different types of shoes that interested me. However, all the reviews were biased because most people had strong views on the product, both good and bad, which influenced their say in the review. Also, the shoes were on many different websites, so the demographics for each website may be different and some people of different demographics may have a different point of view than someone from a different demographic. A signal problem existed because I was aware of how opinions of non-professionals are not a good indicator of how to judge a product, idea, or anything else. In order to counteract this signal problem, I would have to have professional shoe raters test the shoe and give feedback on the shoe. If their input was similar to the customers’ reviews, I know that it is a trustworthy decision to either buy/not buy depending on the result. If their input was different, I would side with the professionals slightly more, but I would average the two scores and see the potential result from that.

  • Profile photo of Tyler McCaw

    The most important part about our discussion of bias was how it can alter people’s decisions in bigger ways than I thought. Also I thought it was interesting to find the bias in online sites like Yelp. Also the signal problem was interesting because I learned you can’t just accept the data as always being correct. You actually have to dig and do research in order to make sure the data you are looking at is accurate. A recent decision I made was getting a new laptop for Christmas. I had to look at many reviews online in order to make sure I was getting a good laptop. Most people are bias when they create a review so I had to be aware of that and try and see if I could find more accurate reviews. I used several websites in order to make my decision which allowed me to asses the reviews clearly and see which ones were potential bias or not.

  • Profile photo of Max Snyder

    The most important takeaway from the bias and signal problem for me was that you should always be skeptical with the data that you are using. An example of this is when I was working on a project for another class on government surveillance. If I had just looked at what data the government was collecting using conspiracy nut websites I would probably have concluded that the government was always watching me. However by going to the NSA’s official website I was able to pull off a full list of the data that the NSA collects from its citizens. In addition to using the NSA’s website I also used some of the articles that we read in this class to get a better idea of what data was really being collected. The key point here is to use data from multiple sources so you will not inadvertently express a bias.

  • Profile photo of Jamya L Day

    The most important takeaway is how not all data is complete or representative of everything/everyone. I’ve recently used Yelp to decide on a restaurant. I had a bias on price, so some restaurants were automatically eliminated as an option because they were too expensive. I possibly could have eliminated really good restaurants because of my price bias. Instead, I could have still considered the more expensive restaurants.

  • Profile photo of Keerthana Rachamadugu

    The most important takeaway from last week’s discussion about bias and signal problem would be that when collecting data or implementing a change, we must really be aware and a conscious effort to eliminate bias and signal problem as much as we can. When I was applying to colleges about a year ago I used a website called Niche to see my chances of getting accepted into my desired colleges and to figure out what colleges I needed to really focus on. Niche had data of students SAT scores and GPA and whether they got accepted, rejected, or waitlisted for that specific college. There were many biases in website where the sample size was too small, they didn’t take into account any other information, and much of the information was probably only from students who really cared and were really happy or sad about the results. There was a signal problem where there wasn’t representation of many different types of people, only the selective amount that used the website. I realized that I should be using a more reliable source and ultimately no website can truly tell me if I would get into my desired colleges because there are way too many variables colleges look at during admission process.

  • Profile photo of Elijah Wilder

    The most important takeaway from last week’s discussion was that bias can be both deliberate and accidental. That being said, I also know how important it is for an individual to properly analyze data in order to avoid bias, and if the data is biased, then the individual must find a new data set. Recently, I was looking at average scores for SATs by state, and I found that the southern states averaged higher than the northern states, but that made no sense since the northern states’ education was better than that of the southern states. That’s when I realized the data didn’t account for the number of student who took the SATs. I had to go to a different data set that showed me that variable, and I found out that even though the northern states had lower score, they all had more students (>50%) who were able to take the test. In some of the southern states, the percentages were in single digits (<10%). The bias made it appear as if states with good standards for education were in need of help, when in fact they were not.

  • Profile photo of Kahleem M Wilkins

    Recently I have gotten a new phone and I used data to help me decide which phone to get. After looking at reviews from multiple sights, I finally decided on the S7 edge due to positive feedback and its specs appealed to me. I was already pretty biased towards the phone before searching other alternatives but the reviews seal the decision for me.

  • Profile photo of Johnny Luu

    Throughout the whole discussion of bias and signal problems in data, the main thing that stuck with me is that most of the data we use is biased in some way. Whether if it is reviews of food, places, events, or products the information found there will be skewed in some type of way. It isn’t always necessarily biased purposely, but still the subjective opinion is present. For example, last week I was trying to decide between two pairs of shoes to by for myself. I looked up reviews of both and got mixed reviews on which shoe to choose. The problem I noticed was that the reviewers factored in other shoes’ quality and design along with the current shoe so it through everything off. I ended choosing by a flip of the coin. My best advice on how to counteract the inevitable bias in the data we use is to take it all “with a grain of salt”. We should all just try to experience or try it our ourselves so we will know if it works for us or not. An opinion from another person always makes me iffy until I can make my own judgement.

  • Profile photo of Alison Kline

    I feel as though the most important takeaway from last week’s discussion on bias is the fact that bias is always there and as long as we are aware of said bias, we can reduce the chance of making a mistake or influencing conclusions when analyzing data. I used the data of various prices when deciding whether to buy Ben and Jerry’s or 7/11 brand ice cream. The Ben and Jerry’s was more expensive, and with my bias, I usually tend to go for the cheaper option when purchasing something. However, my bias of knowing that I like Ben and Jerry’s and being loyal to the brand, also had an influence over my hesitation in automatically buying the cheaper ice cream. There could have been a signal problem because there were only two brands of ice cream to choose from, not all brands of ice cream were there for me to base my decision on. I might have counteracted this bias by attempting to base my decision on the physical ice cream rather than by the price or my brand bias.

  • Profile photo of Sarah Trested

    In my opinion, the most important takeaway from last week was the topic of data fundamentalism which is the notion that correlation always indicates causation, and that massive data sets and predictive analytics always reflect objective truth. There hasn’t really been a recent case in which I collected data, however, the first example I thought of for this questions is in regards to the recent presidential election. Many people, based on data and statistics, believed that Hillary Clinton would win the election. The reason why the data was incorrect and Donald Trump won the election, is likely to be due to an error in the collection of data is regards to data bias and a signal problem.

  • Profile photo of Parth B Patel

    My biggest takeaway from last week’s class on bias and signal problem is being aware of the data. The data that you have or you obtain should be questioned to check for its integrity. We must as the “why” and “how” questions to provide greater higher level insight into that results. Also, data must be checked to see if it represents the population or the sample size and for accuracy. A quick example of me using data is choosing NFL(football) players for fantasy sports line up on Fanduel. The concept is fairly simple, look at the various players and their past data in regards to their performance from previous weeks and versus other teams. Then select players who I think will play well in order to win. A lot of times I run into a dilemma with bias choosing my players from my favourite team, even though the numbers might say otherwise. To deter any bias or signal problems I can pick games in which my favourite team is not playing, which will eliminate personal preferences.

  • Profile photo of Emma Girandola

    During the discussion on Monday I found it to be the most interesting that large data sets can be greatly skewed due to the signal problem. Data can be skewed due to the fact that in low income areas people cannot afford smartphones where data is taken from. Therefore, large portions of the population are left out of data sets. Recently I was online shopping for my best friend’s 21st birthday and when I was looking through the Forever 21 website I was having a very difficult time finding something that I would like to buy. The main reason being that when I looked at the pictures of the model wearing the clothing I thought the article of clothing was very pretty. But then when I read the reviews they were very extreme. For the most part the only reviews that were there were reviews that were highly praising the article of clothing or saying how terrible they were. These reviews could be affected by the signal problem if the people that bought the clothing do not have access to a computer or smart phone. Some articles of clothing that are sold online are also sold in stores. The reviews of the clothing that are also sold in the stores are not being accounted for.

  • Profile photo of Jibreel Murray

    In my opinion, the most important takeaway from the discussion regarding bias and the signal problem is that biases are inevitable in Big Data and the signal problem will always under-represent the population. I recently use data when deciding what classes I would take for this semester. I used the Rate my Professor website when picking classes. The data on the professors researched are biased, with the ratings going from 1 to 5 and mostly students bashing their professor if they did not get a good grade. A signal problem in the data on rate my professor is that all of the students who have taken the professor do not leave a rating. This skews the data making it bias to students who only have had a great or terrible experience with the professors.

  • Profile photo of Amy Huynh

    In my opinion, the most important takeaway from last week’s discussion on bias and the signal problem is to always be conscious/aware of data and to analyze it. Bias is widespread, and it can make decision-making difficult. Moreover, it is important to thoroughly analyze data and question it instead of simply accepting it. An example of how I used data to make a decision was when I was debating on whether or not I wanted to buy a specific make-up product. I became interested in the product after watching a video on a make-up artist rave about it. However, when I read reviews on the product, there were people who either loved it or disliked it which wasn’t very helpful. Reading the reviews left me conflicted and skeptical, and I didn’t know whether or not I wanted to purchase it. To counteract the bias and signal problem, I decided to buy the make-up product and try it out since I’m able to return it if I’m not satisfied.

  • Profile photo of Robin Goetz

    For me, the most important takeaway from our discussion regarding bias and the signal problem is that biases are inevitable when working with data. Humans are always involved when collecting data. They can be the ones that provide the data or the ones that designed a way to collect the data. It is human to be biased, humans create data, thus data will always be biased, too. It is important to remember that once you are aware of the bias you can take action to reduce it and secure the integrity of you insights won from data. I recently used data to decide which smartphone I would buy. I looked into different manufacturers and operating systems and ended up buying a Samsung smartphone. One bias was that I wanted to buy an Android phone, since I am learning how to build apps for that OS, and my previous negative experience with Sony. Although I also looked at Apple and Windows phones, I was more looking to find reasons why to get a Samsung instead of actually comparing the different phones’ technical data. Therefore, you can definitely say that there was a bias present, which also resulted in a Signal Problem. To counteract this signal problem I could have copied only the technical data in a word file and remove the name and brand of the phone. Then look for the best data and match it with the phone. I would have made a decision based on solely data and not have been influenced by emotions.

  • Profile photo of Ellen Gbeyeh Wiah

    The most important takeaway from last week class was acknowledging that bias exist in data collecting and interpretation and that the signal problem can affect the credibility of the data. Over the weekend I was decided where to order food, I look at reviews on yelp and google reviews and decided on Italian Kitchen a local pizzeria up the street from my house. I picked that pizzeria based on the review and on the fact that I have order from there before and the food was good, unlike zesto’s where I had a bad experience in the past. Due to my bad experience at zesto’s a bias already existed in my head that their service was bad before reading the reviews.When dealing with bias it’s best to collect more data, increase the sample size to minimize errors. It also helps to be specific about the kind of data you are looking for.

  • One of the most interesting things about last week’s discussion was the fact that we never are able to completely eradicate bias from the data. I am really interested in this specific aspect since I have a big interest in the stock market and we can often see how stock indices are biased, for example, in the Dow Jones Average indices are biased towards high priced values by representing one share of the company in the index. I do not feel as if the sample affected the entire analysis so I do not believe that there was a signal problem.

  • The take away from last week’s lesson was that bias is inevitable when taking data, also it is important to keep in mind that sampling error leads to a for-sure discrepancy in the population being sampled and the whole it is being applied to. Every day people make decisions: What to eat, where to go, what to read, these choices may be between one or two things. For instance I can choose to read my history book or a novel I am reading on the side. This seems as if it is a binary choice, read one book or the other, the truth is that I could choose to read any book. The bias of my life has lead me to only consider these two choices.

  • Profile photo of Patrick Murt

    The most important takeaway from last week’s class in my opinion was the identification of personal bias as a signal problem in data collection. An example of bias would be how I was looking for places to eat this weekend on Zomato, but for almost all of the restaurants I saw had mixed reviews, either boasting about good food, or complaining about bad food. These reviews weren’t reliable because only people who experienced great or poor meals at the restaurant would be more likely to leave reviews, as oppose to people who’ve had average meals. In the end, I just ended up getting McDonald’s and calling it a night.

  • We mostly talk about bias and Signal Problem on last week’s discussion. The most important takeaway in the discussion I think is the bias and Signal Problem can’t be removed completely from the data sets even people use large data pool. Even thought what we do will reduce the errors, but it’s inescapable that people will still have some errors. In order to prevent the error happen on us, it takes steps to eliminate bias from the data sets. Rate my professor is a website that helps colleges’ students to decide if they are take the class or not base on the evaluations from professors’ students. One of the biggest bias we can usually see is sometimes the rates of the same professor can be 1.0 and 5.0 at the same time. How could this happen? For some of the students who didn’t get a good grade from the, they can still “grade” their professor. Those evaluations give wrong information for viewers.

  • Profile photo of Luke Greenley

    The most important thing that I took away following reflection on biases in data sets and the Signal Problem is the simple realization that data is not neutral. As the article by Kate Crawford put it, they are creations of human design. We are the ones that point out relationships between data sets, and we decide what they mean. Recently, I used data sets to make a decision on signing a lease for an apartment. I compared data and performed cost-benefit analysis for each apartment with my prospective roommates, and we came to a decision. However, the biases I had during this process included my feelings towards a certain unit and the features of the unit such as price, location, size, amenities, etc. I do not believe a signal problem arose from these biases though, because I was not the only one making the decision on the apartment so compromise with my roommates was a factor. Also, the comparisons to other units prevented a blind assumption or inference on an apartment, especially having toured the units previously. Regardless, bias in data sets is definitely something to keep in mind for the future when comparing data, as it has possible major implications for an outcome.

  • Profile photo of Moeed Faisal

    My most important takeaway from last time’s discussion was the skepticism we always need when approaching data sets. Although we may have reduced error to a minimum, there is always something that can be questioned whether that be the sampling method, sample size, biases, or the Signal Problem. I recently used some data to find the best bagel to buy at Bagel Hut. I looked at the options and gathered information by asking customers what their favorite bagel was and why that was so. I found that most people at that time recommended me to get the spicy cheddar with cream cheese. I realize that there was definitely bias in my little sampling. For starters, what I did was a convenience sample, I just asked the people available at the time. Second, this was at 9 o clock in the morning, there is probably a difference in tastes between those who come early vs those who come later. Also, this wasn’t a very representative sample of the Bagel Hut customer base. When I usually come, there are always professors, Caucasians, African Americans, Middle Easterners, men and women. That day however, there were mostly older professors all Caucasian. I obviously had a biased sample. The signal problem also applies to this situation, I had lost the voices of other people who didn’t have class in the morning and/or those who woke up later. In the end, however, the spicy cheddar was delicious.

  • Profile photo of Malik Lassiter

    The most important takeaway I had from the class discussion was that of the signal problem. I still think it is strange that the bias of the signal problem can’t be removed from the situation no mater how big the data pools are. I recently relied on Amazon revews to make decision for a purchase for a hair product and i can see how bias would be encountered there since it sounded like the customers who left a review were either very satisfied with the product or completely disappointed with and there were not many people in the middle. I could have tried to counteract the bias by checking the item/s reviews on other websites or trying to ask my family and friends if they had ever had used it before.

  • I believe the most important takeaway from last week’s discussion was that although data is a powerful tool, it is never the complete answer. Useful research will always take this into account and supplement findings with other measures for accuracy and completeness. For example, when I was using census data to make conclusions about different hypotheses for the first assignment, I most likely would not reach the most accurate conclusions about how the income per capita of a city is representative of the average wealth of a city’s resident’s. This is because I would probably encounter a biased finding from a data set easily affected by a signal problem. In this case, census data can create a signal problem if measures are not taken to correct for underrepresentation of people with income below the poverty line, who have many factors that prevent them from fully partaking in census data collection. I might have accounted for this signal problem by supplementing my findings with specific measures of poverty to cross-check the accuracy of my initial conclusions.

  • Profile photo of Jordan Massey

    The biggest takeaway for me was the fact that bias truly cannot be eliminated from data, it can be minimized but not taken out. It made me think of all the statistics and polls we hear about in the news and how we have to take information with a grain of salt because even statistics have heavy bias in them.

  • My most recent use of Data was in a Sports Marketing class, We used statistics of venues to determine prices, ticketing, and promotion

  • Profile photo of Giovanni Mancini

    One takeaway I garnered from this weeks lessons’ is that bias is within every study by virtue of the fact that human beings are biased, and human beings conduct studies. All though I was not specifically using the data, the dissenting opinions of the new presidents cabinet as well as specifically from himself about the validity of polls from left-wing sources is something I find interesting. First of all, the election polls were wrong from many outlets, but many of them used the popular vote rather than by checking on a state by state basis. A new myth is pervading political discourse, that is that statistical studies can be drawn up in any way based on the bias of the person creating the analysis. I do not know if this is true, but what I do know is that although there is bias within every study and news outlet, that alone should not be grounds to denounce anyone who disagrees with you. In fact, data and studies contradicting your opinions should be the ones you pay the most attention to.

  • Here’s what I took out of last week’s class discussion. If you have a signal problem or any bias information then you have to take the right steps to get rid of the bias datasets. It’s important if you have any signal problems that you figure out the problems before it actually occurs. We see this so often in so many datasets. In my most recent data that I used for my sports radio show, we were talking about the Super Bowl and asked people to take a poll on if it was the best Super Bowl ever played. In our data we got back people that it was the best comeback, but not the best Super Bowl ever. Now, obviously there is a lot of biasness towards this poll and dataset because some people don’t like giving the Patriots credit and some people don’t like them as a team. This is something we see in sports all the time that’s why data in sports are very hard to come by, because a lot of it is opinionated.

  • Profile photo of Timothy Goebig

    The most important thing I took out of last week’s discussion on bias and signal problem is to actually assess the data that is produced right in front of me. My most recent incoming with this situation is picking my fantasy basketball team for the week. I never thoroughly looked at stats that the players on my team provide, but rather just picked on the performance from previous games. I know now to look at stats, and imply them into game time situations and how they play against the team at hand.

  • Profile photo of Jordan Massey

    The most recent instance of the use of data in my every day life is looking at the first week sales of some of my favorite music artists to determine their popularity. I have determined that if an artist exceeds a first week sales number of over 50k, they have reached a new increment of popularity and 100k and so on and so on. I believe that the one thing that we can look at in music to quantify popularity is the first week sales because that is when demand for the the artist’s project is the highest.

  • The most important thing to take away from this would be that if there is a signal problem the most important thing is to find it and stop it before it corrupts data. one instance in which used data was to figure out what movie i wanted to watch next. I took ratings, expert reviews and run time into account when choosing the movie, this is something that I have never done before and after doing it i found it very helpful because it gave me more information about them and influenced my choice.

  • Profile photo of Nicholas Papagno

    The most important thing to take away from last week’s class discussion is there is no way to completely get rid bias data. For example, it is hard to watch any news station if you are neutral in politics. NBC tends to lean towards the Democrats meanwhile FOX tends to lean towards the republicans. Therefore most infographics, stories, and opinions are all going to make their side look a bit better. Even on election night, the different news stations were contradicting each other. Therefore, bias data is based off the humans that input it or display it, making it difficult for a viewer to truly understand what it happening.

Leave a Reply

Your email address will not be published. Required fields are marked *

Office Hours
Ermira Zifla (instructor) 10:00am-12:00pm Wednesdays, Speakman Hall 207C or by appointment.
Prince Patel (ITA) by appointment only. Email:
Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 20 other subscribers