Sentiment
Sentiment
Correlation
I created a lower correlation matrix of all the fashion topic terms to observe the correlation of sentiment.
Some things to note in the table below. We can see there is a lack of deeper green colours so I have noted limes. Gingham seems to correlate with a positive sentiment with country fashion, blouses, and the colour pink. Other fashion topics with a positive correlation include sapphire, purple, and eyeliner. Interestingly sweater and stripes correlate.
On the other hand, negative sentiment correlations include season terms summer and winter, specifically colours. Summer had a negative correlation with stripes, eyeliner, and the colour black. Whilst winter had a negative relationship with red and blue.
High correlation | Low correlation |
---|---|
lipstick & brown | summer & brown |
purple & eyeliner | heels & red |
purple & yellow | summer & stripes |
blouse & country | winter & red |
sapphire & purple | winter & blue |
gingham & country | eyeliner & summer |
sapphire & blouse | summer & black |
stripes & sweater | |
gingham & blouse | |
pink & gingham |
Below dives in deeper with two terms with correlated highly with others, sweater and purple. We can see that sweater has a high positive correlation with stripes, cardigan, dating, and pink (0.6<x).
On the other hand, purple correlates with yellow, eyeliner, and sapphire (0.7<x), but correlates the most low with summer (-0.1>x)
Sweater | Purple |
---|---|
Sentiment
The above boxplot shows sentiment analysis across of fashion terms with deeper insight into the plots below.
Compared to other fashion accessories, the ring is lower sentiment whilst a bracelet is higher. Eyeliner is a high sentiment on average compared to mascara (lower). In patterns: despite stripes having a positive correlation with sweater, it is on average lower sentiment. Finally, heels and boots have a lower sentiment as footwear compared to pumps, trainers, and sandles.
End of page.