Prologue: In a Data Shire Far Away in Copenhagen
In the quiet Shire of Data Science, where rows and columns lived in harmony, there existed many visualization tools, akin to the peaceful hobbits of this land. But amidst these, a tool emerged - HiPlot, the One Ring to rule them all, that did not just display data but unveiled layers of hidden insights.
ℹ️ One Ring used in a !=
👿 way. Code and data behind my PyData Copenhagen Talk that inspired this article
Chapter 1: The Fellowship of the Tools
Like Frodo Baggins, a humble hobbit thrust into an unexpected journey, I ventured into the vast downland woods of data matching toward forging models that could predict apartment sale price with the fellowship of mat plotting tools.
Mat plotting tools at my disposal were akin to the trusted, yet simple, tools of a Hobbit's life. Seaborn
, for instance, was a loyal companion. Its ability to generate statistical data visualisation to lit correlation matrices, distribution and outliers with violin or box plots were invaluable.
While useful, their process was akin to walking the long, winding paths of Middle Earth on foot. Generating correlations or identifying outliers was informative, but not swift.
In the fast-paced world of the Dark Lord of Mordor, feature selection and model results diagnostics, where decisions and iterations need to happen at the speed of thought, these tools, like a slow-moving ent, lacked the urgency required.
Chapter 2: The Eye of Sauron - Spotting Outliers
Then, as Gandalf presented the One Ring to Frodo, HiPlot came into my grasp. Like the swift steeds of Rohan - fast, powerful, and majestic. HiPlot transcended the traditional methods, offering a visualization experience that was not just informative but also incredibly rapid. Its interactive and intuitive interface allowed for a faster, almost instantaneous exploration of data, aligning perfectly with the needs of quick-thinking ML development.
Armed with HiPlot, I confronted the Eye of Sauron - outliers in my datasets. Like Frodo using the ring to see the unseen, HiPlot illuminated these anomalies in stark contrast against the norm.
Each outlier stood out, clear as the Eye atop Barad-dûr, no longer a hidden threat to my model’s integrity.
Chapter 3: The Paths of the Dead - Unearthing Features
In the Paths of the Dead, Frodo had to trust his instincts to navigate. Similarly, HiPlot guided me through the spectral world of feature selection. It revealed which variables were mere echoes and which were vital to my quest of predicting apartment prices.
Like the Army of the Dead aiding Aragorn, these features became my allies.
The final leg of my journey was akin to Frodo's ascent of Mount Doom. Understanding the inner workings of my models was a treacherous path, fraught with complexity and confusion. Yet, HiPlot, like Samwise Gamgee, was my steadfast companion, shedding light on the strengths and weaknesses within my algorithms.
Palantír of R2 and RMSE in One Future
Together, we unraveled the mysteries, bringing clarity to chaos. We connected dots of the proportion of the variance for a dependent variable that is explained by independent variables, R2
by selecting the range of RMSE
.
Epilogue: HiPlot - The Ring of Power
As Frodo's journey transformed him from a simple hobbit to the savior of Middle Earth, HiPlot transformed my approach to data. It wasn’t just a visualization tool; it was a source of truth and insight, a guide through the perilous realm of data science. In the land of models and metrics, HiPlot stands as the One Plot to rule them all, a beacon of clarity in an often-convoluted world.
Until
then, keep on coding with the Horn of Rohan
Top comments (2)
That was a great talk!
Thank you @shahnoza 💡If one learnt 2/8 ideas I shared then my task was completed :D