Instruction to QBUS6860 Assignment 2
We strongly suggest you consider Assignment 2 as a research project and expect you finally present a research paper alike report. You can refer to many research papers on their pattern or style and understand what information is expected in such a report.
For your convenience and our expectation, I am attaching a sample report from 2021S2 QBUS6860. Although the topic is different from ours for this semester, but it does serve as a good example that you can follow.
Basically your report shall contain the following information:
Copyright By PowCoder代写 加微信 powcoder
(a) A title
(b) Background story (telling your audience or readers what it is about)
(c) Your question to be answered or your hypothesis to be verified in the project and
its meaning and importance (that is about why and your motivation for the project). It is important for you to choose a topic first, e.g., who is the leading country in AI/Machine Learning research, how other countries can improve etc.
(d) Your methodology (that is how you solve it to get answers/insights/conclusions) and toolsets etc
(e) Data description (what facts you rely on and their formats, or how you change the data for your purpose.)
(f) Results (could be presented visually with explanation)
(g) Your insights for Machine Learning community (this is not about a business) (h) What can be further improved and your suggestions if any
(i) References
Note: The University new policy asks only your ID for all written assessments. Please insert your ID in the header area.
1. Background ………………………………………………………………………………………….2 2. Business Question and Justification ……………………………………………………….2 3. Methodology, Hypothesis and Justification of Selected Analytical Tools ……….3 4. The Data ……………………………………………………………………………………………….6 5. The result ………………………………………………………………………………………………7 6. Insights for Business ……………………………………………………………………………………….9 7. Limitations ……………………………………………………………………………………………10 8. References …………………………………………………………………………………………….10
QBUS6860 Individual Assignment 2 Page 2 of 18
Your Student ID
Due date: Student ID:
QBUS6860 Visual Data Analytics (2022S1)
Individual Assignment 2
Monday 23 May 2022 ______________________________________________
Your Student ID
The Impact of Big Data Analytics on Innovation in Formula 1 1. Background
The challenges and opportunities posed by Big Data have attracted the attention of both scholars and industry practitioners (Akhtar et al., 2017; Marr, 2015; Reinmoeller and Ansari, 2016). Academic research acknowledges the opportunities offered by Big Data when information (data) is translated into decision making strategies and improved innovation and performance (Chen, Chiang, and Storey, 2012; McAfee and Brynjolfsson, 2012; Mc Institute, 2011). Current debate also points to the competences companies need to deal with advanced technology and Big Data (Akhtar et al., 2017; Reinmoeller and Ansari, 2016). Although the benefits and challenges identified in the management literature are numerous, the link between Big Data and innovation remains largely anecdotal due to lack of empirical work on how large datasets can influence business outcomes (Kache and Seuring, 2017; Sen, Ozturk, and Vayvay, 2016). Several scholars have proposed frameworks for how Big Data applications can be exploited to generate value, relying mainly on a case-based research methodology (Matthias et al., 2017). However, the generalizability of these findings to a wider population of companies is difficult. The lessons learnt may be unique to the in-situ performance at a particular time. A systematic review conducted by Frizzo-Barker et al. (2016), of Big Data papers published between 2009 and 2014, acknowledges the lack of empirical work and shows that this stream of research is ominated by conceptual papers.
2. Business Question and Question Justification
This report addresses the following business question: what is the impact of Big Data analytics on innovation in Formula 1? This question is not only applicable to the Formula 1 industry, but also to other analytically dense industries. Yet, F1 constitutes an ideal setting to investigate the use of Big Data (Aversa, Cbantous, and Haefliger, 2016; George, Haas, and Pentland, 2014). First, in the F1 context, major innovations are distinctly and precisely measured (Gino and Pisano, 2011). Furthermore, unlike many industries (where products/services are heterogeneous), all F1 teams produce homogeneous output (final standing in the race ranking), which allows us to compare team performance directly and more precisely (Goodall and Pogrebna, 2015). Second, F1 is a highly analytically dependent and data-dense industry and has seen a transition from systematic manual processing of data to predictive analytics and, more recently, evolution to a mature stage in the Big Data revolution. Their performance depends on how the teams respond to these data, which makes a good testbed for information processing theory (Rogers, Miller, and Judge, 1999; Tushman and Nadler, 1978) operationalized in the Big Data operations management domain. Third, in most high technology industries, the strictness of safety and regulatory standards increases over time as more information is revealed. Examples are the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) and Restriction of Hazardous Substances (RoHS) regulation, which is aimed at tighter control of EU supply chains by monitoring substances used in products (Westervelt, 2012). This tighter regulatory control involves increasingly more complex data collected from manufacturers. The F1 industry has experienced the imposition of many regulations over time (Marino et al., 2015) and how teams respond and adapt to process ‘future data’ into existing systems before making an
QBUS6860 Individual Assignment 2 Page 3 of 18
Your Student ID
informed decision about the design of cars and race strategies provide important lessons for other industries.
Scholars have recognized the distinctive characteristics of F1, for example, Jenkins (2010) suggests that F1 firms possess sustaining capabilities – munificent resource configurations that extend the time available for firms to adapt to technological changes – thereby allowing them to remain competitive across discontinuities. In this study, I focus on the Big Data structure (BDS) of F1 teams from the perspectives of information integration and information accessibility. Building on an information processing perspective, which emphasizes the need to process information by considering external demand (Rogers et al., 1999; Tushman and Nadler, 1978) and the technological environment in which firms operate (Hughes et al., 2014; Hughes, Hughes, and Morgan, 2007), I examine how different strategies for Big Data information processing lead to different BDS configurations. I show how different BDS structures influence team performance in terms of output (achievement of podium positions) as well as innovation production.
3. Methodology, Hypothesis and Justification of Selected Analytical Tools
F1 is recognized as a unique setting to investigate the role of Big Data and business analytics, since it relies heavily on sophisticated applications of real-time information systems to support informed decision making processes during a race (Aversa et al., 2016; Aversa, Furnari, and Haefliger, 2015; George et al., 2014; Goodall and Pogrebna, 2015; Marino et al., 2015). F1 is estimated to be worth approximately $6 billion annually (Sylt and Reid, 2011). Constructor teams’ profits come from advertising and TV. A higher finishing position, primarily a podium position (1st to 3rd), generates more sponsorship and TV income. Increasingly, modern teams are raising money from the development of F1 technologies that spill over into other industries. For example, Williams and McLaren (Applied Technologies) have associate companies. It is an interesting industry intellectually because it is subject to a great deal of regulatory turbulence. The Fédération Internationale de l’Automobile (FIA), the F1 industry governing body, imposes strict conditions, which are revised annually, on all aspects of F1 (the teams, technology, resources, track, tyres, drivers, etc.). The link between regulation and innovation has been well documented (Stewart, 2010). T is embodied in F1; regulation is unambiguously associated with innovation and performance (Jenkins, 2004, 2010; Jenkins, Pasternak, and West, 2007; Khanna, Kartik, and Lane, 2003; Marino et al., 2015) and regulatory compliance results in a level playing field for all competing teams. Indeed, sometimes rule changes are made with the specific intention of curtailing the dominance of one team, for example, Ferrari and in 2003 (Hoisl, Gruber, and Conti, 2017).
F1 is an extremely data-dense industry with sophisticated data analytics. All contemporary F1 cars are using sophisticated telemetry systems to obtain, transmit, process, and analyse information. According to NASA, the term telemetry originates from the Greek “tele” which means “remote” and “metron” which refers to “measure” and depicts a process of automatized communications and transmissions allowing to obtain data from remote or poorly accessible points for monitoring (SAO/NASA Astrophysics Data System, 1987). In the F1 context, telemetry is implemented through a large number of sensors and electronic
QBUS6860 Individual Assignment 2 Page 4 of 18
Your Student ID
devices inclusive of the Electronic Control Unit (ECU) which communicates data to the pit wall, pit garage or another remote site (e.g., Toet, 2013).
Each Formula 1 car is equipped with 150-300 sensors (dependent on the racetrack, weather conditions, and other factors). For the reasons explained below, F1 telemetry is a one-way transmission system: the data is sent from the car to the engineering and strategy team, but the team does not have an opportunity to send the data to the car. According to various technical forums1, the data from the car is transmitted wirelessly using 1,000-2,000 encrypted telemetry channels using either 1.5 GHz frequency or another locally allowed frequency. While the delay between the data collected and received at the team boxes varies, on average, it is around 2 milliseconds. Since the received data is compressed, the number of actual gigabits of data received by teams may differ from race to race, although each team collects approximately 1.5 billion samples of data from a single race and approximately 5 billion of samples throughout the race weekend (this includes data from all training sessions). The transmitted data on engine performance, suspension state, gearbox performance, fuel status, temperature readings including tires temperature, g-forces and actuation of controls by the driver is analysed by the engineering and strategy team and results of this complex live analytics is communicated to the driver in the form of racing strategy advice. Since telemetry is the major source of (live) Big Data for the F1 teams, in my analysis I use the telemetry development stages in the F1 industry as a “natural” proxy of (Big) data analytics evolution. Since 1950s, an F1 team performance was highly dependent on this team’s ability to collect, process, and analyse large amounts of data. Historically, we can distinguish between 6 phases in the F1 data analytics progress using the development of telemetry as a proxy for determining the boundaries between different stages. The Big Data analytics history of the F1 industry proxied through the evolution of the use of telemetry was drawn from Jenkins (2010), the data collected by the McLaren F1 team2 as well as from the Formula 1 Dictionary Technical Forum.3
I distinguish between the following phases in F1 data analytics (depicted in Table 1). The evolutionary phase-based approach summarised in Table 1 allows us: (i) to identify the pretelemetry period (Phase 1); (ii) to identify phases of significant heterogeneity between F1 teams in terms of their access to telemetric technology (Phase 2 and Phase 4); as well as (iii) to understand when F1 teams had similar or standardized telemetric technology (Phases 3, 5, and 6). Using the identified phases, I can now use F1 performance data to understand the differences between phases and explore whether performance of industry as a whole as well as performance of individual players changed from one phase to the other.
To develop my hypothesis about innovation in Formula 1, I use a combination of Linden and Fenn (2003) and Fenn and LeHong (2011) Cycle framework. I assume that analytics phases shown in Table 1 determine the boundaries of Cycle stages of innovation development: Technology Trigger, Peak of Inflated Expectations; Through of Disillusionment; Slope of Enlightenment; and Plateau. Therefore, my main hypothesis is that the lifecycle of major innovations in Formula 1 follow the Cycle shape where each stage is determined by the data analytics phase from Table 1. Specifically, I hypothesise that the correspondence between analytics and innovation will follow the pattern depicted on Figure 1. Specifically, Phase 1 (Driver as a Sensor) pre-dates telemetry analytics and, therefore, represents a preliminary stage. This stage leads to Phase 2 (Telemetry
QBUS6860 Individual Assignment 2 Page 5 of 18
Your Student ID
Development) where the development of telemetry technology is triggered which causes as increase in the number of innovations. Phase 3 (Early Telemetry Phase) represents the peak of inflated expectations where data potential is uncovered by the industry and positive “hype” is created allowing to reach the peak of innovation. Phase 4 (Turbulence) captures a “through of disillusionment” where drawbacks of the telemetry technology start to significantly outweigh the benefits and number of innovations rapidly decrease. Phase 5 (Mature Telemetry) phase is equivalent to the slope of enlightenment where new capabilities of technology provide a new positive boost to innovation. This boost flattens or even disappears in Phase 6 (Big Data Telemetry) when technology reaches its plateau or even post plateau stage.
Table 1: Phases of Data Analytics in F1 Industry Phase Time period Major milestones
1. Driver as a Sensor
The majority of teams used drivers as sensors who fed back the information about the car performance to the teams after the race.
2. Telemetry Development
1975 – McLaren started to experiment with telemetry first deployed 14 sensors on IndyCar. Until late 1980s – F1 teams started to experiment with
telemetry.
3. Early Telemetry
By 1989 – F1 teams used “patched” telemetry transmitting data when cars came close to pits Early 1990s – F1 teams had high rate live information but it had blind spots (especially on tracks with dense trees or high buildings like Monza, Monaco, etc.). So information was incomplete.
1998 – Plextek4 became a major supplier for telemetry systems
2000 – Incomplete information problem was fixed
4. Turbulence
2002 – Two-way telemetry was allowed (teams could not only receive but also send information to cars remotely)
2003 – FIA banned two-way telemetry
5. Mature Telemetry
2005 – Electronic Control Unit (ECU) TAG-310B SECU by McLaren Electronic Systems and Microsoft is developed
2008 – FIA standardized ECU for all F1 cars
2008 – FIA makes Advanced Telemetry Linked Analysis System (ATLAS) Express produced by McLaren Electronic Systems standard
6. Big Data Telemetry
2013 – Volumes of data become very high requiring major system upgrades
2013 – Upgraded standard ECU to the new TAG-320 SECU
QBUS6860 Individual Assignment 2 Page 6 of 18
Your Student ID
2017 – Data Viewer developed by McLaren Applied Technologies5
Why wouldn’t Big Data telemetry provide an extra boost to the innovation lifecycle in Formula 1 instead leading to a plateau? Information Processing Theory (IPT) provides a theoretical basis for this (Rogers et al., 1999; Tushman and Nadler, 1978). Early research on IPT (Daft and Lengel, 1986; Daft and Weick, 1984) maintains that information gaps can be reduced by gathering more data. A large part of IPT discusses the reduction of uncertainty by facilitating decision makers’ access to the right information at the appropriate time (Sakka, Barki, and Côté, 2016). However, obtaining more data is today not a major concern since most electronic devices and transactions generate abundant data. At the same time, greater access to data is linked to higher levels of misinformation and misinterpretation of those data. Although all F1 teams have access to Big Data, this does not mean that all of them will necessarily benefit from the data. In fact, greater volumes of information can be very difficult to process so only few market players may be capable of coping with the industry’s increasing data supply.
Figure 1: The Major Innovations Lifecycle in the Formula 1 Industry Predicted Based on the Phases of Data Analytics Evolution
In my analysis, I am planning to capture the major innovations’ lifecycle using bar chart and provide several F1 teams use case illustrations using line chart. Instead of a single line chart, I am planning to use a chart with multiple lines, because it is more appropriate to demonstrate the differences between teams. Though bar chart is a good way to demonstrate life cycle, I will also use area chart in order to capture the dynamic shape better.
4. The Data
Data used in this analysis spans 68 years from 1950 to 2017. The data includes 976 F1 races that took place in that period, with 22,083 car records forming our dataset. For each car record, I have data on: the starting and final positions of the cars that participated in each
QBUS6860 Individual Assignment 2 Page 7 of 18
Your Student ID
race, that is, team performance; constructor teams; their leaders’ names, personal information and background; drivers’ personal information and background; and information on each race circuit (weather conditions on the day, length of the circuit, etc.). The data were compiled from several sources (please, see my dataset attached as a separate file). For information on car entries, circuit, constructor, drivers and other detailed Grand Prix race information, I used the provided dataset. Records in the starting dataset were cross-checked and augmented by other sources of data such as Wikipedia, Grand Prix Encyclopaedia, and other websites.6 Data on F1 regulations were compiled from the website of the F1 regulatory body, Federation Nationale d’Automobile, https://www.fia.com/regulations.
Main dependent variable
My dependent variable is radical innovation. Radical innovation variable captures significant changes in technology which led to serious changes in the F1 industry (innovations associated with telemetry are excluded). In order to construct this variable, I used FIA regulations which were scanned for restrictions on or bans of major innovations complemented by data from a broad variety of technical online forums.7 My dataset also includes the performance variable (final standings at the end of the race) for every race car in every Grand Prix season since the first year of the Formula 1 industry existence. I use the final race standings ranking as a determinant of performance for each car with 0 identifying the race winner; 1 – 2nd
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com