Foundations — Publication-quality figures — LaTeX labels, colormaps, DPI

5.4.19 · D1 · Coding › Scientific Computing (Python) › Publication-quality figures — LaTeX labels, colormaps, DPI

Koi bhi dial ghumaane se pehle hume har ek word aur symbol ka matlab agree karna hoga. Parent note mein figsize, dpi, points, pixels, viridis, r"...", $...$ jaisi cheezein bhari hain. Ek samajhdaar 12-saal ka baccha inme se koi nahi jaanta. Isliye hum har ek cheez ko zero se banate hain, uss order mein jahan har naya idea sirf pehle wale ideas par tikta hai.

1. Inch — woh ruler jiske against sab kuch measure hota hai

Picture (Figure 1): neechey ek ruler khicha hai jo inch scale dikhata hai; amber bar exactly ek inch mark karta hai, aur ek chota caption yaad dilata hai ki wo 2.54 cm ke barabar hai. Uske upar, ek cyan rectangle ek 3.5-inch-wide figure hai jo us ruler par baitha hai, ek double-headed cyan arrow uski 3.5 in width measure karta hai. Abhi computers ka koi chakkar nahi — yeh woh length hai jo tum haath se measure kar sakte ho.

Topic ko yeh kyun chahiye: journals "badi" ya "choti" figures nahi maangti — wo exact column width in inches maangti hain (jaise 3.5 in). Is topic mein baaki har measurement (points, DPI) inch ke relative define hoti hai. Yeh anchor hai. Yeh miss karo toh baaki kuch bhi meaningful nahi rehta.

2. Pixels aur dots — sabse chhote specks (aur kyun wo alag hain)

Picture (Figure 2): left panel mein ek smooth cyan sine curve hai jo bahut saare chhote pixels se bani hai; right panel mein wahi curve sirf nau mote amber pixels se rebuild ki gayi hai (asli curve white mein ghost ke roop mein peechhe hai), isliye wo blocky lagti hai — Minecraft ki tarah. Same shape, far fewer specks per inch.

Topic ko yeh kyun chahiye: "sharpness" ka matlab hai bahut saare chhote specks per inch; "blocky" ka matlab hai kuch bade specks per inch. Sharpness ki baat karne ke liye pehle speck ka word chahiye — aur upar wale PPI/DPI trap se bachne ke liye.

3. DPI — ek inch mein specks kitne tight pack hain

Ab hum idea 1 (inch) aur idea 2 (pixel/dot) ko jodte hain.

"Per inch" kyun, na ki "total dots"? Kyunki tumhari aankhein picture ko paper par uski physical size par padhti hain. Ek million dots ek wall par failaye toh blocky lagte hain; wohi million dots ek stamp mein squeeze karo toh razor-sharp lagte hain. Sirf dots per inch hi pakad paata hai ki wo kitna sharp lagta hai. Exactly isliye unit inches se divide karta hai — dekho Dimensional analysis.

4. figsize — canvas ki physical size

Picture (Figure 3, left side): ek wide cyan rectangle jis par "default 6.4 in canvas" likha hai versus ek narrower amber rectangle jis par "3.5 in column" likha hai, dono par double-headed arrow apni width deta hai. Dono par identical word "Intensity" likha hai — notice karo wo wider card ke relative chhota lagta hai. Woh relationship is pure topic ka dil hai.

Topic ko yeh kyun chahiye: journal ki column width hi tumhara figsize width hai. Ise pehle set karo, hamesha.

5. Point — woh unit jisme fonts measure hote hain

Picture (Figure 3, in full): cyan 6.4-in canvas par ek bada, padhne laayak "Intensity" hai; amber 3.5-in canvas par wohi 9-pt label shrink-to-fit ke baad chhota render hua dikhta hai. Neechey white caption shrink factor $3.5/6.4$ spell out karta hai.

figsize font size se pehle kyun set karein? Kyunki point inch se tied hai, isliye inch canvas pehle fix hona chahiye; tabhi "9 pt" printed page par woh matlab rakhta hai jo tum sochte ho. Parent ki rule "figsize first, always" ka reasoning yehi hai.

6. Raster vs vector — drawing store karne ke do tarike

Topic ko yeh kyun chahiye: line plots (axes, curves, text) shapes hain → vector PDF ke roop mein save karo, infinitely sharp. Photos aur dense heatmaps pehle se colors ki grids hain → high DPI par raster PNG ke roop mein save karo. Zyada jaankari Raster vs vector graphics mein.

7. String, raw string, aur $...$

Ab text dial ki baat. Do symbols sabko confuse karte hain: leading r aur dollar signs.

Picture (Figure 4): code r"$\lambda$" bada dikhaya gaya hai. Ek cyan arrow r ki taraf point karta hai note ke saath "keep the backslash (Python side)"; ek amber arrow $...$ ki taraf note ke saath "turn on math mode (LaTeX side)"; neechey, clean italic result $\lambda$ appear hota hai. Do arrows, do alag kaam — dono zaroori hain.

8. mathtext vs text.usetex

Choice kyun matter karti hai: agar tumhara paper Computer Modern font use karta hai, toh usetex figure text ko seamlessly blend in karta hai; mathtext kaafi close aata hai lekin identical nahi.

9. Colormap aur perceptual uniformity

Topic ko yeh kyun chahiye: ek honest figure mein woh features nahi honay chahiye jo data mein hain hi nahi. Ek monotonic (hamesha-brighten-hone wala) colormap tumhari aankhon ko sach rakhta hai.

10. rcParams — har dial set karne ki ek jagah

Prerequisite map

Yeh map kaise padhein (Figure 5): arrows prerequisite se us cheez ki taraf point karte hain jo wo enable karta hai. Inch (amber, top) woh root hai jisse sab kuch ugta hai; cyan arrows neechey follow karo aur sab amber goal box "Publication-quality figure" par milte hain bottom mein.

Map top-down goal ki taraf padhta hai: inch DPI, figsize aur point ko feed karta hai; point aur figsize milke font legibility decide karte hain; strings math mode ko feed karti hain; aur sab kuch publication-quality figure mein funnel ho jaata hai.

Equipment checklist

Answers cover karo aur khud test karo — tum topic ke liye ready ho agar har ek baat bol sako.

Connections

• Parent topic
• Matplotlib basics — figure and axes objects
• Colormaps and color theory in visualization
• LaTeX typesetting
• Raster vs vector graphics
• Dimensional analysis
• Reproducible research and rcParams