In this final assignment - the grand finale for BCB 520 - I am polishing up one of the figures in my thesis/manuscript… stacked bar charts!
THE ORIGINAL VERSION
What do you think?
Barrie pointed out that it would be helpful to display the random missing samples from my dataset. So, I added in the missing samples (originally not there) and went through the process again.
THE POLISHED VERSION
Do you see the difference?
Barrie was right! This is easier to follow. And shocker, my advisor likes it much better too.
IN CONCLUSION
Sometimes what seems “unimportant” actually is important. Grateful for this class and all it continues to teach me, data-viz-wise and otherwise.
MORE FIGURUES
Just for fun! Here’s a few examples for how I simplified the data.
JaccDiet
JaccWeek
Source Code
---title: "Final Assignment - Polishing Up My Figures"subtitle: "Stacked bar chart fun"format: html: toc: false echo: trueauthor: "Heidi Sellmann"date: "2024-04-27"categories: [Portfolio, DataViz, Microbiome, Bar charts, Assignment]image: "grandFinale.jpg"description: "Missing data is not meaningless!"code-fold: truecode-tools: truecode-link: trueeditor: markdown: wrap: 72---```{r}knitr::opts_chunk$set(echo=FALSE, warning=FALSE, error=FALSE, message=FALSE)```# OVERVIEWIn this final assignment - the grand finale for BCB 520 - I am polishing up one of the figures in my thesis/manuscript... stacked bar charts!```{r Load libraries, warning = FALSE, message=FALSE, output = FALSE}library(phyloseq);packageVersion("phyloseq")library(tidyverse);packageVersion("tidyverse")library(vegan);packageVersion("vegan")library(reshape2);packageVersion("reshape2")library(magrittr);packageVersion("magrittr")library(microbiome);packageVersion("microbiome")``````{r Load data orig}genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG <-readRDS("genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG.rds")genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG <-readRDS("genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG.rds")```# THE ORIGINAL VERSION```{r Create color palette HM and IF}#First create an initial 30 color palette to do the initial plotslibrary(RColorBrewer)custom_col30_C <-c("#FF0000", "#00B0F0", "#FFFF00", "#96D050", "#CC3399","#375623", "#FFC000", "#0070C0", "#990033","#00B050","#FF00FF", "#66FF99", "#F96E05", "#FFFF99", "#000000","#0000FF", "#FF7C80", "#CC66FF", "#00FF00", "#002060","#5F5F5F", "#FF0066", "#666633", "#FF99FF", "#CCCC00","#66FFFF", "#660033", "#D9D9D9", "#666699", "#660066")``````{r Create top 20 genera plot HM and IF, echo=FALSE, output = FALSE}#HMtop20generaPlotHM <-ggplot(genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x") +# facet_grid(TRMT~.) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("HM Top 20 Genera")top20generaPlotHM#IFtop20generaPlotIF <-ggplot(genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x") +# facet_grid(TRMT~.) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("IF Top 20 Genera")top20generaPlotIF#HM# Extract the digits after the underscore in Sample_IDgenus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG <- genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG %>%mutate(Sample_ID =sub("^\\d+_([A-Za-z]+\\d+)$", "\\1", Sample_ID)) %>%mutate(Sample_ID =factor(Sample_ID, levels =unique(Sample_ID[order(as.numeric(sub("^\\D*(\\d+).*", "\\1", Sample_ID)))])))# this is the same from above! # Create the plot with the ordered x-axistop20generaPlotHM <-ggplot(genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x", labeller =labeller(Week =function(x) paste("Week", x))) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("HM Top 20 Genera")top20generaPlotHM# using lebeller above to add Week label!#IF# Extract the digits after the underscore in Sample_IDgenus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG <- genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG %>%mutate(Sample_ID =sub("^\\d+_([A-Za-z]+\\d+)$", "\\1", Sample_ID)) %>%mutate(Sample_ID =factor(Sample_ID, levels =unique(Sample_ID[order(as.numeric(sub("^\\D*(\\d+).*", "\\1", Sample_ID)))])))# this is the same from above! # Create the plot with the ordered x-axistop20generaPlotIF <-ggplot(genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x", labeller =labeller(Week =function(x) paste("Week", x))) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("IF Top 20 Genera")top20generaPlotIF# using lebeller above to add Week label!```## What do you think? Barrie pointed out that it would be helpful to display the random missing samples from my dataset. So, I added in the missing samples (originally not there) and went through the process again. # THE POLISHED VERSION```{r Load data polished}genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG_0s <-readRDS("genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG_0s.rds")genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG_0s <-readRDS("genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG_0s.rds")``````{r HM 0s plot, warning=FALSE, echo=FALSE, output = FALSE}#HMtop20generaPlotHM0s <-ggplot(genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG_0s, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x") +# facet_grid(TRMT~.) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("HM Top 20 Genera")top20generaPlotHM0s#HM# Extract the digits after the underscore in Sample_IDgenus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG_0s <- genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG_0s %>%mutate(Sample_ID =sub("^\\d+_([A-Za-z]+\\d+)$", "\\1", Sample_ID)) %>%mutate(Sample_ID =factor(Sample_ID, levels =unique(Sample_ID[order(as.numeric(sub("^\\D*(\\d+).*", "\\1", Sample_ID)))])))# this is the same from above! # Create the plot with the ordered x-axistop20generaPlotHM0s <-ggplot(genus_glom_rel_HM_data_clean_wide_ordered_Top20_LONG_0s, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x", labeller =labeller(Week =function(x) paste("Week", x))) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("Top 20 Genera in HM Group")top20generaPlotHM0s# using lebeller above to add Week label!``````{r IF 0s plot, warning=FALSE, echo=FALSE, output = FALSE}#IFtop20generaPlotIF0s <-ggplot(genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG_0s, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x") +# facet_grid(TRMT~.) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("IF Top 20 Genera")top20generaPlotIF0s#IF# Extract the digits after the underscore in Sample_IDgenus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG_0s <- genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG_0s %>%mutate(Sample_ID =sub("^\\d+_([A-Za-z]+\\d+)$", "\\1", Sample_ID)) %>%mutate(Sample_ID =factor(Sample_ID, levels =unique(Sample_ID[order(as.numeric(sub("^\\D*(\\d+).*", "\\1", Sample_ID)))])))# this is the same from above! # Create the plot with the ordered x-axistop20generaPlotIF0s <-ggplot(genus_glom_rel_IF_data_clean_wide_ordered_Top20_LONG_0s, aes(x = Sample_ID, y = Abundance, fill=Top20genus)) +facet_wrap(Diet~Week, scales ="free_x", labeller =labeller(Week =function(x) paste("Week", x))) +geom_bar(stat ="identity") +scale_fill_manual(values=custom_col30_C) +theme(axis.text.x =element_text(angle =60, hjust =1)) +ggtitle("Top 20 Genera in IF Group")top20generaPlotIF0s# using lebeller above to add Week label!```## Do you see the difference? Barrie was right! This is easier to follow. And shocker, my advisor likes it much better too. # IN CONCLUSIONSometimes what seems "unimportant" actually is important. Grateful for this class and all it continues to teach me, data-viz-wise and otherwise.# MORE FIGURUESJust for fun! Here's a few examples for how I simplified the data. 
