SIT 214 – SCIENTIFIC RESEARCH METHODS

What is the Scientific Method?

The scientific method is a process for experimentation that is used to explore observations and
answer questions. Does this mean all scientists follow exactly this process? No. Some areas of
science can be more easily tested than others. For example, scientists studying how stars change as
they age or how dinosaurs digested their food cannot fast-forward a star's life by a million years or
run medical exams on feeding dinosaurs to test their hypotheses. When direct experimentation is
not possible, scientists modify the scientific method. In fact, there are probably as many versions of
the scientific method as there are scientists! But even when modified, the goal remains the same: to
discover cause and effect relationships by asking questions, carefully gathering and examining the
evidence, and seeing if all the available information can be combined in to a logical answer.
Even though we show the scientific method as a series of steps, keep in mind that new information
or thinking might cause a scientist to back up and repeat steps at any point during the process. A
process like the scientific method that involves such backing up and repeating is called an iterative
process.
Whether it is a science fair project, a classroom science activity, independent research, or any other
hands-on science inquiry understanding the steps of the scientific method will help focus the
scientific question and work through the observations and data to answer the question as well as
possible.
Steps of the Scientific Method
1. Ask a Question
The scientific method starts when a question is asked about something that is observed: How, How
many, What, When, Who, Which, Why, or Where?

2. Do Background Research
Rather than starting from scratch in putting together a plan for answering the question, be a savvy
scientist using library and Internet research to find the best way to do things and ensure that
mistakes from the past are not repeated.

3. Construct a Hypothesis
A hypothesis is an educated guess about how things work. It is an attempt to answer your question
with an explanation that can be tested. A good hypothesis allows to then make a prediction:
"If _____[I do this] _____, then _____[this]_____ will happen."
State both the hypothesis and the resulting prediction to be tested. Predictions must be easy to
measure.

4. Test the Hypothesis by Doing an Experiment

The experiment tests whether the prediction is accurate and thus the hypothesis is supported or not.
It is important for the experiment to be a fair test. A test done fairly by making sure that only one
factor is changed at a time while keeping all other conditions the same.
The experiments should be repeated several times to make sure that the first results weren't just an
accident.

5. Analyze the Data and Draw a Conclusion

Once the experiment is complete, the measurements are collected and analyzed to see if they
support your hypothesis or not.
Scientists often find that their predictions were not accurate and their hypothesis was not supported,
and in such cases they will communicate the results of their experiment and then go back and
construct a new hypothesis and prediction based on the information they learned during their
experiment. This starts much of the process of the scientific method over again. Even if they find
that their hypothesis was supported, they may want to test it again in a new way.

6. Communicate the Results

To complete the scientific project, the results will be communicated to others in a final report and/or
a display board. Professional scientists do almost exactly the same thing by publishing their final
report in a scientific journal or by presenting their results on a poster or during a talk at a scientific
meeting.

Data Science
Data science is an inter-disciplinary field that uses scientific methods, processes, algorithms and
systems to extract knowledge and insights from many structural and unstructured data. Data science
is related to Data Mining and Big data.
Data science is a "concept to unify statistics, data analytics, machine learning and their related
methods" in order to "understand and analyse actual phenomena" with data. It employs techniques
and theories drawn from many fields within the context of mathematics, statistics, Computer
Science, and Information Science. Turing Award winner Jim Gray imagined data science as a
"fourth paradigm" of science (empirical, theoretical, computational and now data-driven) and
asserted that "everything about science is changing because of the impact of information
technology" and the data deluge.

Data Science Project Life Cycle

Given the right data, Data Science can be used to solve problems ranging from fraud detection and
smart farming to predicting climate change and heart diseases. With that being said, data isn’t
enough to solve a problem, you need an approach or a method that will give you the most accurate
results. This brings us to the question:

How Do You Solve Data Science Problems?

A problem statement in Data Science can be solved by following the below steps:
1. Define Problem Statement/ Business Requirement
2. Data Collection
3. Data Cleaning
4. Data Exploration & Analysis
5. Data Modelling
6. Deployment & Optimization

Data Science Project Life Cycle – Data Science Projects – Edureka

Let’s look at each of these steps in detail:
Step 1: Define Problem Statement
Before you even begin a Data Science project, you must define the problem you’re trying to solve.
At this stage, you should be clear with the objectives of your project.
Step 2: Data Collection
Like the name suggests at this stage you must acquire all the data needed to solve the problem.
Collecting data is not very easy because most of the time you won’t find data sitting in a database,
waiting for you. Instead, you’ll have to go out, do some research and collect the data or scrape it
from the internet.
Step 3: Data Cleaning
If you ask a Data Scientist what their least favorite process in Data Science is, they’re most
probably going to tell you that it is Data Cleaning. Data cleaning is the process of removing
redundant, missing, duplicate and unnecessary data. This stage is considered to be one of the most
time-consuming stages in Data Science. However, in order to prevent wrongful predictions, it is
important to get rid of any inconsistencies in the data.
Step 4: Data Analysis and Exploration
Once you’re done cleaning the data, it is time to get the inner Sherlock Holmes out. At this stage in
a Data Science life-cycle, you must detect patterns and trends in the data. This is where you retrieve
useful insights and study the behavior of the data. At the end of this stage, you must start to form
hypotheses about your data and the problem you are tackling.
Step 5: Data Modelling
This stage is all about building a model that best solves your problem. A model can be a Machine
Learning Algorithm that is trained and tested using the data. This stage always begins with a
process called Data Splicing, where you split your entire data set into two proportions. One for
training the model (training data set) and the other for testing the efficiency of the model (testing
data set).
This is followed by building the model by using the training data set and finally evaluating the
model by using the test data set.
Step 6: Optimization and Deployment:
This is the last stage of the Data Science life-cycle. At this stage, you must try to improve the
efficiency of the data model, so that it can make more accurate predictions. The end goal is to
deploy the model into production or production-like environment for final user acceptance. The
users must validate the performance of the models and if there are any issues with the model then
they must be fixed in this stage.
Now that you know how a problem can be solved using Data Science, l
https://www.edureka.co/blog/data-science-projects/

Why Data Science?

Why? Because over the past few years, companies have been storing their data. And this being done
by each and every company, it has suddenly led to data explosion. Data has become the most
abundant thing today.
But, what is to be done with this data? Let’s understand this using an example:
A company which makes mobile phones releases a first product, and it became a massive hit. Every
technology has a life, right? So, now it’s time to come up with something new, but the company
don’t know what should be innovated, so as to meet the expectations of the users, who are eagerly
waiting for the next release?
Somebody, in the company comes up with an idea of using the user generated feedback and pick
things which the users may be expecting in the next release. Comes in Data Science, one applies
various data mining techniques like sentiment analysis etc... and gets the desired results. It’s not
only this, the company can make better decisions, can reduce its production costs by coming out
with efficient ways, and give its customers what they actually want! With this, there are countless
benefits that Data Science can result in, and hence it has become absolutely necessary for the
company to have a Data Science Team.

The term Data Science has emerged recently with the evolution of mathematical statistics and data
analysis. The journey has been amazing, we have accomplished so much today in the field of Data
Science.
In the next few years, we will be able to predict the future as claimed by researchers from MIT.
They already have reached a milestone in predicting the future, with their awesome research. They
can now predict what will happen in the next scene of a movie, with their machine! How? Well it
might be a little complex for you to understand as of now, but don’t worry by the end of this blog,
you shall have an answer to that as well.
Coming back, we were talking about Data Science, it is also known as data driven science, which
makes use of scientific methods, processes and systems to extract knowledge or insights from data
in various forms, i.e either structured or unstructured.
What are these methods and processes, is what we are going to discuss in this Data Science Tutorial
today.
Moving forward, who does all this brain storming, or who practices Data Science? A Data
Scientist.
Who is a Data Scientist?

As you can see in the image, a Data Scientist is the master of all trades! He should be proficient in
maths, he should be axing the Business field, and should have great Computer Science skills as
well. Scared? Don’t be. Though you need to be good in all these fields, but even if you aren’t,
you’re not alone! There is no such thing as “a complete data scientist”. If we talk about working in a
corporate environment, the work is distributed among teams, wherein each team has their own
expertise. But the thing is, you should be proficient in at least one of these fields.

How to solve a problem in Data Science?

So now, let’s discuss how should one approach a problem and solve it with data science. Problems
in Data Science are solved using Algorithms. But, the biggest thing to judge is which algorithm to
use and when to use it?
Basically there are 5 kinds of problems which you can face in data science.

Let’s address each of these questions and the associated algorithms one by one:
Is this A or B?
With this question, we are referring to problems which have a categorical answer, as in
problems which have a fixed solution, the answer could either be a yes or a no, 1 or 0, interested,
maybe or not interested.
For Example:
Q. What will you have, Tea or Coffee?
Here, you cannot say you would want a coke! Since the question only offers tea or coffee, and
hence you may answer one of these only.
When we have only two type of answers i.e yes or no, 1 or 0, it is called 2 – Class Classification.
With more than two options, it is called Multi Class Classification.
Concluding, whenever you come across questions, the answer to which is categorical, in Data
Science you will be solving these problems using Classification Algorithms.
Is this weird?
Questions like these deal with patterns and can be solved using Anomaly Detection algorithms.
For Example:
Try associating the problem “is this weird?” to this diagram,

What is weird in the above pattern? The red guy, isn’t it?
Whenever there is a break in pattern, the algorithm flags that particular event for us to review. A real
world application of this algorithm has been implemented by Credit Card companies where in, any
unusual transaction by a user is flagged for review. Hence implementing security and reducing
human’s effort on surveillance.
How much or How many?
Those of you, who don’t like maths, be relieved! Regression algorithms are here!
So, whenever there is a problem which may ask for figures or numerical values, we solve it using
Regression Algorithms.
For Example:
What will be the temperature for tomorrow?
Since we expect a numeric value in the response to this problem, we will solve it using Regression
Algorithms.
Moving along in this Data Science Tutorial, let’s discuss the next algorithm,
How is this organised?
Say you have some data, now you don’t have any idea, how to make sense out of this data. Hence
the question, how is this organised?
Well, you can solve it using clustering algorithms. How do they solve these problems? Let’s see:

Clustering algorithms group the data in terms of characteristics which are common. For example in
the above diagram, the dots are organised based on colors. Similarly, be it any data, clustering
algorithms try to apprehend what is common between them and hence “clusters” them together.
What should I do next?
Whenever you encounter a problem, wherein your computer has to make a decision based on the
training that you have given it, it involves Reinforcement Algorithms.
For Example:
Your temperature control system, when it has to decide whether it should lower the temperature of
the room, or increase it.
How do these algorithms work?
These algorithms are based on human psychology. We like being appreciated right? Computers
implement these algorithms, and expect being appreciated when being trained. How? Let’s see.
Rather than teaching the computer what to do, you let it decide what to do, and at the end of that
action, you give either a positive or a negative feedback. Hence, rather than defining what is right
and what is wrong in your system, you let your system “decide” what to do, and in the end give a
feedback.

Data Science Training

Let’s apply this understanding in the example above, imagine you are training the temperature
control system, so whenever the no. of people in the room increase, there has to be an action taken
by the system. Either lower the temperature or increase it. Since our system doesn’t understand
anything, it takes a random decision, let’s suppose, it increases the temperature. Therefore, you give
a negative feedback. With this, the computer understands whenever the number of people increase
in the room, never increase the temperature.
Similarly for other actions, you shall give feedback. With each feedback your system is learning and
hence becomes more accurate in its next decision, this type of learning is called Reinforcement
Learning.
Now, the algorithms that we learnt above in this Data Science Tutorial involve a common “learning
practice”. We are making the machine learn right?
What is Machine Learning?

It is a type of Artificial Intelligence that makes the computers capable of learning on their own i.e
without explicitly being programmed. With machine learning, machines can update their own code,
whenever they come across a new situation.
Concluding in this Data Science Tutorial, we now know Data Science is backed by Machine
Learning and its algorithms for its analysis. How we do the analysis, where do we do it. Data
Science further has some components which aids us in addressing all these questions.
Before that let me answer how MIT can predict the future, because I think you guys might be able
to relate it now. So, researchers in MIT trained their model with movies and the computers learnt
how humans respond, or how do they act before doing an action.
For example, when you are about shake hands with someone you take your hand out of your pocket,
or maybe lean in on the person. Basically there is a “pre action” attached to every thing we do. The
computer with the help of movies was trained on these “pre actions”. And by observing more and
more movies, their computers were then able to predict what the character’s next action could be.
Easy ain’t it? Let me throw one more question at you then in this Data Science Tutorial! Which
algorithm of Machine Learning they must have implemented in this?

Operation Research: Applications, Methodology and Tools

Definition of Operation Research:
Uniform and acceptable definition of O.R. is not available.
However, few definitions available in literature are given below:
1. “Operation research is a scientific method of providing executive departments with a quantitative
basis for decisions regarding the operations under their control”.
2. “Operation research is concerned with scientifically deciding how best to design and operate
man-machine systems usually under conditions requiring the allocation of resources”.
3. “Operation research is a scientific approach to problem solving for executive management”.
4. “Operation research is the application of scientific method by interdisciplinary teams to problems
involving the control of organized (men-machines) systems so as to provide solution which best
serve the purpose of the organisation as a whole”.
5. “Operation research is an aid for the executive in making his decisions by providing him with the
needed quantitative information based on the scientific method of analysis”.
6. “Operation research in the most general sense can be characterized as the application of scientific
methods techniques and tools to problems involving the operations of systems so as provide those
in control of the operation with optimum solutions to the problems”.
Applications of Operation Research:
Operation research is a problem solving and decision taking technique. It is considered a kit of
scientific and programmable rules which provides the management a “quantitative basis” for
decisions concerning the operation under its control.
Some areas of management where Operation research techniques have been successfully
utilized are as follow:
1. Allocation and Distribution in Projects:
(i) Optimal allocation of resources such as men materials machines, time and money to projects.
(ii) Determination and deployment of proper workforce.
(iii) Project scheduling, monitoring and control.
2. Production and Facilities Planning:
(i) Factory size and location decision.
(ii) Estimation of number of facilities required.
(iii) Preparation of forecasts for the various inventory items and computation of economic order
quantities and reorder levels.
(iv) Scheduling and sequencing of production runs by proper allocation of machines.
(v) Transportation loading and unloading,
(vi) Warehouse location decision.
(vii) Maintenance policy decisions.
3. Programmes Decisions:
(i) What, when and how to purchase to minimize procurement cost.
(ii) Bidding and replacement policies.
4. Marketing:
(i) Advertising budget allocation.
(ii) Product introduction timing.
(iii) Selection of advertising media.
(iv) Selection of product mix.
(v) Customer’s preference of size, colour and packaging of various products.
5. Organization Behaviour:
(i) Selection of personnel, determination of retirement age and skills.
(ii) Recruitment policies and assignment of jobs.
(iii) Recruitment of employees.
(iv) Scheduling of training programs.
6. Finance:
(i) Capital requirements, cash flow analysis.
(ii) Credit policies, credit risks etc.
(iii) Investment decision.
(iv) Profit plan for the company.
7. Research and Development:
(i) Product introduction planning.
(ii) Control of R&D projects.
(iii) Determination of areas for research and development.
(iv) Selection of projects and preparation of their budgets.
(v) Reliability and control of development projects thus it may be concluded that operation research
can be widely utilized in management decisions and can also be used as corrective measure.

Methodology of Operation Research:

Quantitative basis for decision making is provided to managers by Operation research. it enhances a
manager’s ability to make long range plans and to solve the routine problems of running a
enterprise/concern Operation research is a systematic and logical approach to provide a rational
footing for taking decisions. Operation research, like scientific research is based on scientific
methodology which involves following steps.
1. Formulating the Problem:
Operation research is a research into the operation of a man machine organisation and must
consider the economics of the operation in formulating a problem for Operation research
study analysis must be made of the following major components:
(i) The environment.
(ii) The objectives.
(iii) The decision maker.
(iv) The alternative courses of action and constraints out of the above four component, environment
is most comprehensive as it provides a setting for the remaining three. The operation researcher
shall attend conferences, pay visits, send observation and perform research work thus succeeds in
getting sufficient date to formulate the problems.
2. Constructing a Model to Represent the System under Study:
Once the project is approved by the management, the next step is to construct a model for the
system under study. The operation researcher can now construct the model to show the relations and
interrelations between a cause and effect or between an action and a reaction.
Now the aim of operation researcher is to develop a model which enables him to forecast the effect
of factors crucial to the solution of given problem. The proposed model may be tested and modified
in order to work under stated environmental constraints. A model may also be modified if the
management is not satisfied by its performance.
3. Deriving Solution from the Model:
A solution may be extracted from a model either by conducting experiments on it i.e., by simulation
or by mathematical analysis. No model will work appropriately if the data is not appropriate. Such
information may be available from the results of experiments or from hunches based on experience.
The data collection can clearly affect the models output significantly. An Operation researcher
should not assume that once he has defined his objective and model, he has achieved his aim of
solving the problem. The required data collection consumes time to prepare if date collection errors
are to be minimized.
4. Testing the Model and the Solution Derived from it:
As has been pointed out earlier a model is never a perfect representation of reality. But if properly
formulated and correctly manipulated, it may be useful in providing/predicting the effect of changes
in control variables on overall system effectiveness.
The usefulness or utility of a model is checked by finding out how well it predicts the effect of these
changes. Such an analysis is usually known as sensitivity analysis. The utility or validity of the
solution can be verified by comparing the results obtained without applying the solution with the
results obtained when it is used.
5. Establishing Controls over the Solution:
The next phase for the operation researcher is to explain his findings to the management. It may be
pointed out that he should specify those conditions under which the solution can be utilized.
He should also point out weaknesses if any so that management will know what risks they are
taking while employing the model to generate results. Thus he should also specify the limits with in
which the results obtained from using the model are valid. He should also define those conditions
under which the model will not work.
6. Implementation of the Solution:
The last phase of the operation research methodology is implementation of solutions obtained in the
previous steps. In operation research though decision making is scientific but its implementation
involves so many behavioural issues. Therefore the implementing authority has to resolve the
behavioural issues. He has to sell the idea of utility of Operation research not only to the workers
but also to superiors.
The distance between Operation research scientist and management may create huddles thus the gap
between one who provides a solution and the other who wants to utilize it must be eliminated. To
achieve this both the management and Operation research scientist should play positive role. A
properly implemented solution obtained through application of Operation research techniques
results in improved working conditions and gains the management support.

Tools of Operation Research:

In any area of human endeavour, whether it is a production system, business system or service
system where an objective is to be optimized, the problem falls into the domain of operation
research.
Some of the commonly used techniques of operation research are as follows:
1. Linear programming.
2. Waiting line theory or queuing theory.
3. Inventory control models.
4. Replacement problems.
5. Network Analysis.
6. Sequencing.
7. Dynamic programming.
8. Assignment problems.
9. Decision theory.
10. Integer Programming.
11. Transportation Problems.
12. Simulation.
13. Goal Programming.
14. Markov Analysis.
15. Game Theory.
16. Heuristic Models.
17. Routing Models.
18. Symbolic logic.