0% found this document useful (0 votes)
17 views4 pages

Descriptive Statistics in Drug Response Analysis

The document outlines an assignment on descriptive statistics, divided into two parts. Part I focuses on basic descriptive statistics, including data analysis of gene expression, viscosity measurements, nitrogen concentration in water, and dollar-guaranies exchange rates, requiring various statistical computations and visualizations. Part II involves exploratory analysis of real data, emphasizing descriptive statistics, data preparation, distribution analysis, correlation analysis, and time series plots.

Uploaded by

Fawaen
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
17 views4 pages

Descriptive Statistics in Drug Response Analysis

The document outlines an assignment on descriptive statistics, divided into two parts. Part I focuses on basic descriptive statistics, including data analysis of gene expression, viscosity measurements, nitrogen concentration in water, and dollar-guaranies exchange rates, requiring various statistical computations and visualizations. Part II involves exploratory analysis of real data, emphasizing descriptive statistics, data preparation, distribution analysis, correlation analysis, and time series plots.

Uploaded by

Fawaen
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Assignment – Descriptive Statistics

Part I – Basic Descriptive Statistics


1. An article in Nature Genetics (Vol. 34(1), 2003, pp.85–90), “Treatment-Specific
Changes in Gene Expression Discriminate in vivo Drug Response in Human Leukemia
Cells,” study the gene expression as a function of a medicine consumption. They
developed an experimental design where the subjects were divided into 2 groups.
One group received a high dose of the drug while the control group received no
medicine. Expression data (measures of gene activity) from one gene are shown in
the following table:

Sample 1 2 3 4 5 6 7 8 9 10 11
High Dose 16.1 134.9 52.7 14.4 124.3 99 24.3 16.3 15.2 47.7 12.9
Control 297.1 491.8 1332.9 1172 1482.7 335.4 528.9 24.1 545.2 92.9 337.1

Sample 12 13 14 15 16 17 18 19 20 21 22
High Dose 72.7 126.7 46.4 60.3 23.5 43.6 79.4 38 58.2 26.5 25.1
Control 102.3 255.1 100.5 160 168 95.2 132.5 443 15.8 176 131.1

a) Compute the sample mean, median and 10% trimmed mean of the data. Also, compute
the variance, standard deviation, range and IQR. Comment on the results.
b) Compute the following plots (by hand) and BRIEFLY describe their important
characteristics. Consider that the plots should allow comparison among groups and
should be as informative as possible.
- Dot plot
- Boxplot
- Stem-and-leaf diagram (Leaf Units=1. No subdivisions).
- Histogram.
- Q-Q plot, comparing with a normal distribution.

2. The following data are the viscosity measurements for a chemical product observed
hourly.
Time 1 2 3 4 5 6 7 8 9 10
Viscosity 47.9 48.8 48.6 43.2 43 47.9 48.1 48 43 42.8

Time 11 12 13 14 15 16 17 18 19 20
Viscosity 48.6 48.3 47.9 43.5 43.1 48 47.2 48.3 43.1 43.2

a. Build (by hand) a time series plot of the data with a marginal dot plot.
b. Analyse the plot. Do you find any interesting features?
3. Excessive nitrogen in water might affect the water usability (in agriculture, domestic
use, ecological use, etc). The government of Victoria faces this issue in the Curdies
River in a somewhat cyclical pattern, and suspect it has to do with the water level.
Then, they investigated the nitrogen concentration and the water discharge,
resulting in the following results:

a) What do the red dot in the plot means?


b) What can you conclude from this plot?

After looking at the plot, the scientist decided to build a new plot, but instead of using the
nitrogen data, they used the natural logarithm of the nitrogen, which results in the following
plot:

c) Why the scientists decided to transform the data before plotting?


d) What can you conclude from this plot?
4. The following is a time-series plot with a marginal histogram. It was built with the
change dollar-guaranies from 2020 to 2021:

a) Analyze the histogram. Describe qualitatively the central tendency, spread and skewness
of the data. According ONLY to the histogram, what is the most probable dollar value at
any time?
b) Analyze the time-series. Analyze general tendencies and deviation from those
tendencies. According ONLY to the time series, what is the most probable dollar value in
2023?
c) Do you reach to the same conclusion using the histogram and the time-series? If not,
what seems to be the issue?
d) A classmate (student 1) considered that the dollar value may change on the different
days of the week. Then, he prepared a boxplot of the dollar value in each day. According
to his result (See Boxplot 1), does the previous hypothesis seems reasonable?
e) Another classmate (student 2) had the same hypothesis. To proof his hypothesis, he
found the average dollar value in each week (Avr) and computed the difference of the
dollar value in each day of the week and with the average dollar value in that week
(Mon-Avr, Tue-Avr, Wen-Avr, …). He also do a boxplot of his result (See Boxplot 2). Do
you see any difference? If so, what is the cause of that difference?

Boxplot 1 Boxplot 2
Part II – Exploratory Analysis of Real Data
5. Use descriptive statistics of the dollar-guaranies change from 2015 to 2024. That is,
give a general description of the data, as well as a practical interpretation (see
question examples in “Appendix B”). To do so, make sure you use the plots
mentioned in “5.b”, “5.c”, and “5.d”. You are allowed to use any software to solve
this exercise (excel, Matlab, R-Studio, python compilator, etc)
a) Prepare the data
- Download metadata from reliable resources. Some optional sources are given in
“Appendix A”
- Label each value with its date (for example, Aug-19-2024) and with its day of the week
(for example, Monday).
b) Perform an analysis of the distribution of the data in each day of the week (Monday,
Tuesday, Wednesday, Thursday and Friday). To do so, use histograms, boxplots, empirical
CDF and Q-Q plot. Compare and contrast the usefulness of the different plots.
c) Use scatter plots and the Pearson Correlation Coefficient to perform a correlation
analysis of the dollar-guaranies change with a variable of your choice (such as the PIB
per capita given in “Appendix A”). Make sure you have AT LEAST 7 data points. If further
processing of the data was required before doing the scatter plot, BRIEFLY justify your
data pre-processing. This might include the use of estimated points (annual average,
interpolated values, etc), omission of points, etc.
d) Use time series plots to analyse the change of the price of dollars/guaranies over time.

Appendix A
- [Link] Dolar-Compra
- [Link] Tipo de
Operación: Compra
- [Link] many different data
- [Link] PIB per Capita
- [Link] many different data
- [Link] Inflación

Appendix B
The following are example questions you might want to answer. It is only to better understand
what “general description” and “practical interpretation” means in this the question 5. You might
not want to answer all the questions given. Moreover, you might find that there are interesting
question not listed in here that you might want to answer.
1. General description.
- Is there any noticeable trend?
- How does the distribution looks like?
- What are some hypothesis of the phenomena that drives the data?
- Does pandemic seems to have any effect on the results?
- Others
2. Practical interpretation
- Is it advisable to save in guaranies or in dollars?
- If you are going to do change operations, how do you know when should you do so?
- Is it advisable to operate (buy or sell) in guaranies or in dollars?
- Others.

Common questions

Powered by AI

Different plot types, such as histograms, scatter plots, boxplots, and Q-Q plots, provide varied perspectives on data. Histograms show overall distribution; scatter plots reveal correlations; boxplots summarise central tendency and variability; Q-Q plots test normality. Using multiple plots helps capture the complexity of data, uncover hidden patterns and relationships, and support multifaceted interpretations, vital in financial data analysis .

Visual data representations such as histograms illustrate the frequency distribution of gene expression levels, making it easier to identify patterns such as skewness, bimodality, or normal distribution. Q-Q plots can compare the distribution of expression data to a theoretical normal distribution, allowing evaluation of normality in gene expression levels. These visual tools can highlight differences between experimental groups and reveal underlying data structures, anomalies, or outliers .

The time-series plot reveals temporal trends, seasonal patterns, and deviations over time, while a histogram provides insights into the overall distribution, central tendency, and skewness. Using both views enables a comprehensive understanding of both dynamic behavior across time and static distributional characteristics, crucial for drawing robust conclusions about currency trends .

Time series plots of the viscosity measurements can reveal temporal patterns such as trends, periodicity, and anomalies. Identifying whether the viscosity increases or decreases over time, occurrences of peaks, valleys, or cycles, and constant versus variable viscosity intervals can direct questions about underlying processes affecting chemical stability or product quality .

The pandemic could introduce anomalies or shifts in trends due to economic disruptions, which might be visible as increased volatility, sudden shifts in central tendency, or variance changes in currency data. Hypotheses explaining these changes could include economic policy shifts, altered demand, supply chain disruptions, and variations in global confidence, impacting currency stability and exchange rates .

Transforming data using a natural log can stabilize variance and normalize the data distribution, making relationships between variables more apparent. For nitrogen concentration, this transformation might expose underlying patterns or correlations with water discharge, clarify trends obscured in raw data, and assist in linearizing relationships for more accurate analyses and interpretations .

Data pre-processing such as calculating annual averages helps manage missing values, minimize noise, and standardize data for accurate analysis. In currency exchange rates, pre-processing ensures that comparisons across different timescales make sense by smoothing out short-term volatility and focusing on longer-term trends or relationships when using tools like Pearson Correlation for analysis .

To compare gene expression levels between the two groups mentioned, one can use the sample mean, median, 10% trimmed mean, variance, standard deviation, range, and interquartile range (IQR). Each measure provides different insights: the sample mean offers the average level of expression, while the median provides a central value less affected by outliers. The variance and standard deviation measure the dispersion, and range and IQR describe the spread of data. These statistics enable a comprehensive comparison of central tendency and variability between the high dose and control groups .

Boxplots can visually represent the distribution, central tendency, and dispersion of currency values across different days. Differences in median values, spread (via IQR), and presence of outliers in boxplots, like those analyzed in hypotheses regarding daily currency values, can suggest day-specific patterns or variances, offering evidence whether particular weekdays experience consistent changes .

Factors include volatility, inflation rates, prevailing trends, and economic conditions affecting both currencies. Analysis of historical trends, as shown in time-series or histogram data, informs stability or frequent fluctuations. Evaluating purchasing power parity and economic stability, alongside statistical summaries (e.g., mean, variance), aids in deciding which currency preserves value better in the long term .

You might also like